More speculation, unconfirmed rumors on the future intensification of the war in the Middle East. Maybe Trump will intensify the conflict. Maybe oil flow will stop. Maybe inflation will rip. Maybe everything will burn. Maybe after the mid-terms.

Maybe not.

I’ve been a pretty heavy buyer over the last month, going against the market’s fears on AI slowdown, CapEx, macro headwinds and inflation.

I built myself what I’d call a pretty strong portfolio, filled with AI names that broke out of daily trends and previous lower highs with large net inflows, on positive news and tailwinds.

Today, I am buying a new vertical. A sector no one talks about anymore, overlooked and even hated. And yet… With net inflows, breaking out lower highs on volume, and possibly giving an opportunity for those who’d like an out of the AI trade.

The market has started to diversify lately. This sector seems like a future winner.

Read more

Every piece of bad news seems to be already priced in, the market is holding all-time highs despite any more bad news or comments while any positive ones trigger some buying. I know we all want green dildos, but this market isn’t weak. On the contrary.

This is latent strength, waiting for a trigger.

Over the last months, I expected ranges and volatility, because funds don’t go from “we’re doomed” to “all in” in two days. They go through months of internal talks, risk analysis, macro and fundamental reviews, and more intellectual masturbation before finally starting to reposition themselves.

That’s what happened, a sluggish, up-and-down tape. I think it’s ending.

I’m now fully invested and starting to add margin, slowly, and this article will go through the why, the how and the which, starting by a performance review and ending with concrete execution plans on my holdings and watchlist.

Performance Review

The system continues to perform well and execution improved over the last months. I’ve made some mistakes, corrected them, learned the lessons I had to learn, and followed them to the letter, bringing the portfolio back to ATH a few weeks after Leopold’s liquidation.

• 2025 Returns: +39.5% vs SPY +16.4%

• 2026 YTD Returns: +109.9% vs. SPY +12.24%

• Since inception (Jan-25): +192.9% vs. SPY +30.65%

Extremely satisfied with the performance, but even more by the progress.

My objective with investing and this Substack isn’t to have the largest numbers. I could do better at the casino. But to develop a system that constantly overperform over the long run, without headaches. Learning the capacity to make money is what will differentiate a decades long outperformer to a lucky strike.

Anyone can get lucky once. That’s not what I’m trying to build.

Before starting for real, I would like to redirect you to what might be my best article, released this month. A must read for all investors I believe, new in the game or not. Psychology is too important to ignore.

Now, let’s get going.

Macro

The market has one real concern right now, and I think it’s poppycock.

Let me go through it fast, and remember that macro is only here to be interesting, certainly not to drive investment decisions. It can explain volatility and understanding it can help us hold through it. But that’s about it. No one can anticipate it, and we shouldn’t try.

Oil & Iran

The one real negative, and the one most likely to fix itself.

The conflict traded bombs for sanctions, oil is flowing again but at a higher cost due to higher insurance costs, transfers or rerouting. Not optimal, but improving as we passed from no oil at all to slightly more expensive oil.

The main issue is that higher oil prices feed inflation, and risks higher rate hikes. I was wrong on the last Fed meeting. I expected them to hold and they unanimously voted to hike. So be it, I still made money because the market doesn’t react to the news, it anticipates them.

I called this Fed fear a free buying opportunity. It was.

Bad news will confirm pricing. We’re there already. Good news on the contrary would be completely unexpected. The market isn’t ready for no hike.

The truth is that rates don’t matter anymore, but the market is too used to look at them, like a grown man still dreaming about his prom girlfriend.

No, rates don’t hurt the market anymore. They hurt Americans buying homes and small businesses taking loans, real pain, but they don’t matter to the AI trade besides investors panicking for nothing and older fund managers buying treasuries for a 5% return while most AI companies fund themselves through markets, not banks, and many earn on treasuries. No impact business wise, while stock returns are much more attractive than 5% yield right now.

The US10Y is high. Sure. The market doesn’t care, and shouldn’t.

As for inflation, it’s a household problem, not a market one. Food and bills increase and some suffer, I am not underestimating this. But the market? Everything is a matter of growth. If you grow faster than inflation, you don’t have an issue. If value creation outpaces prices, everything is healthy for a business. If most businesses do, everything is healthy from the market point of view.

Inflation by itself means nothing. It is relative to growth. Everything is relative.

The whole negative macro narrative rests on one single thread: oil. Which itself rests on an even tinier thread: Iran’s war. Ask yourself: is it more likely Trump escalates, or that one deal happens over night?

One signature and this story is over.

Inflation’s here, the Fed hiked, oil’s climbing, Trump won’t make a deal. And the S&P closed at a new high this week. That much bad news should have tanked it long ago. It didn’t. Because there is something even stronger.

Two letters: AI.

The proof is already here

I repeat myself a lot. But that’s how convictions are built, isn’t it?

Fundamentals have never been stronger. Demand either. ROI has no ceiling we can see just yet. CapEx will increase. Pessimism isn’t founded on facts, but opinions.

If you want data, look at Muse.

It hit 5 million downloads in 22 days. ChatGPT took 56. Grok took 103. Claude took 492. This isn't a nerdy code tool, it's a grandma-friendly agent. The first real retail AI agent, and the demand is violent.

Muse is what every company wants to/will offer. A user dedicated agent specialized for a task or a knowledge. Muse is particularly wide and can do almost anything, but some agents will be extremely personalized. Their objective will be to remove even more friction from our daily life, wherever they can find some.

This is the future. But here’s what matters for us today.

Meta, one of the most compute-rich companies on earth, couldn’t meet its demand. Users started to receive limitation notifications passed ~700,000 active users. Why? Not enough hardware. Each user is allocated a dedicated virtual machine with some specific CPU and memory configurations, and compute/data access happens outside of those machines to increase privacy.

They had enough for less than a million users. On a 3.60B daily user pool.

Then OpenAI rushed out its own version, Dots, weeks later. It’ll hit the same wall. Every single retail and enterprises agent will run this wall, by design. Companies built infras to meet training and inference, but they still need the infra to run their service with efficient privacy settings.

Retail AI agents became real. That means more compute, more hardware, more data centers… The trade I’ve been in for a year. This isn’t a Meta story, this is another confirmation that hardware demand has no ceiling we can see, yet.

So yeah.

I understand markets and liquidity mechanisms well enough to know that stocks can’t go up in a straight line. I understand that some are concerned about inflation and CapEx spending. But I also understand that data leans to more spending with access to funding, that inflation impacts Americans and maybe the world through the dollar but not the markets, and that there is only one place where you can make your money work lately, backed by facts, not just Burry’s “I think so” mentality.

That place is the AI trade.

On another subject, If you expect me to comment on Anthropic IPO, I won’t. Numbers are from 2025 so until we have fresh data, anyone talking about it is just looking for attention.

I don’t care about clicks. I care about returns.

So here’s what we do

The same thing we always do in a range: build.

I spent two months buying the strongest names at the best risk/reward possible and then sat on my hands holding winners through fear while cutting bad seeds. We focused on compute and hardware from last year until summer. We started to shift to SaaS and AI applications last month.

Buy strength during the fear. That’s done. Hold winners through the trend. That’s where we are. Trim the extended. That’s where we’re going. Following a simple system. Today’s returns and my fresh ATH reflects what I did two months ago, not yesterday.

December’s ATH will reflect what we’ve been doing until now, and how we’ll behave.

One mistake I’m fixing

I didn’t take profits aggressively enough these last two months. My system is meant to build large positions and lets them run. This is right for uptrends as I won’t cut their wings too early but wrong for a choppy market.

I held AAOI, META and PATH through big gains and watched two of them give it all back, when clear resistances were free trim zones.

In a strong uptrend, you have to let winners run. In a shaky market, trimming into resistance is survival. That would have improved my returns considerably and that’s what I’ll be working on during the next months. Hopefully, we will be back on trends, this is what I expect. But there are always opportunities to improve.

Taking UiPath as an example, my buying and holding were crystal clear on it being a market perform. I didn’t expect it to be my next Soitec. It had what it takes to catch up on its valuation from an underpriced position. I took advantage of this with close to perfect execution but ended up treating it like a Soitec, by not trimming a textbook price action and previous high retest.

From now on, I will manage those positions a bit more aggressively in this kind of context. I’ve already added trim conditions on the FM sheet to force the discipline, and will do more in the next months.

This is what I want my Substack to be. Not a place where I give my buy and sell so it can be copied. But a place where we all learn together how to build and improve a system that works for each of us.

Making money isn’t valuable if it cannot be reproduced.

Now, the specifics

Below, I go through the portfolio position by position; where each stands, what I expect, and exactly how I’ll manage accumulation and trims.

Convictions first, planning second.

Read more

The objective of this article is to share the common lessons we learnt through our years of investing to help new investors or those with lower returns than expected, to up their game, and realize some harsh truth of investing.

Lessons that took us years of mistakes to learn, framed in a small article.

Introduction

For years, my results were mediocre. Not terrible. Mediocre, which is worse, because mediocre lets you believe you’re close. It lets you believe it’s possible, without letting you realize that only drastic changes will allow you to reach your goals.

I was doing what I’d been taught from books and investors who made a fortune in the 1950s. Find good companies. Buy cheap. Hold. Be patient. It took me too long to see the real problem: I wasn’t playing the right game.

Truth is, the world isn’t what it used to be anymore. Markets changed. And the ways to make money in it also changed. But this will only be obvious in the next years, when this decades’ top investors write their own books with their own methods. While those will already be outdated then…

Innovation accelerated, capital allocation changed, liquidity increased, stocks became accessible to everyone, new tools emerged, risk management isn’t the main concern anymore…

Nothing is as it used to be.

The market became a competition to find good stocks, not good companies. Those are widely different concepts, and understanding the gap between them cost me years in mediocre returns.

This piece is written with FJ Research; two investors who made most of the classic mistakes, some less classic ones, corrected them, and landed on the same handful of lessons from different roads. No theory here. Practical lessons, real life experiences and how we actually pivoted to increase our returns.

1. The market isn’t an intelligence contest

The market is a psychological and emotional game where opinions are exchanged for dollars.

Each opinion relies on different objectives, capacities, plans, data sets… No two investors buy for the same reason with the same end goal. Some don’t even know why they buy.

Every data set is valuable, but should never be used in and of itself. Fundamentals and valuation exist for a reason, but if they were enough traders wouldn’t succeed. On the contrary, price action or option flows are filled with information, but if they were enough day traders would be kings. Hunting cheap stocks spending hours readings 10Qs or building better models won’t cut it, just like using complex Ichimoku structures won’t be enough by itself.

Math is useful. Data is important. But don’t underestimate guts.

If it was all formulas, all mathematicians would be billionaires.

Niels Nygaard

TFBTM - I have seen this play both ways.

• Many investors still today argue that Adobe and PayPal are two of the cheapest companies with the largest potential in the market. It might be true, but while most AI hardware names with large narrative rest on 1,000%+ returns the last two years, both are down 22% and 36% over the past years - more on the last five.

• On the contrary, I have personally made a lot of returns on crypto and Strategy. I think we can all agree that both aren’t the strongest fundamentally speaking. And yet, the market bought the assets and pushed their price higher. Greed and emotions rule the market.

How to improve: Treat the market as what it is: a psychological game. There are a thousand ways to win. Find the one that fits you, not the one you read in a book, and most importantly do not dismiss any data set because someone told you this isn’t the right way to invest. Every data set can be used to generate alpha. Find out how.

An investor is judged on one thing: outperformance over years. If he beats the market, he’s good. Regardless of how.

2. No two positions are equal

There is a unilateral truth in the market saying that all positions should follow the same patterns. If you invest on fundamentals, you should only do this and treat all of your positions the same way.

But this isn’t true.

Some come with higher conviction. Some better price action. Some more upside, some less risk. Some can be held decades but others shouldn’t be looked at for more than a month. None are equal, so none should be sized or held equally. Each stock should be bought, held and sold on its own plan.

Know what you own, and know why you own it.

Peter Lynch

TFBTM - My best example would be Hims, one of the rare perfect trades since I write this Substack. I bought it on the start of the GLP-1 compounding rumors and rode it until the first signs of crackdown and deceleration. From the low teens to the $60s. I sold while most social media were still extremely bullish but my thesis was gone, GLP-1 was slowing down - factually, and the market bought this narrative, no others. While many chose to diamond hands, I let it go. I bought with conditions to hold. Conditions broke. I sold.

FJR - I run two types of positions and I treat them differently.

1. Core positions – long-term holdings. I only buy them at a price I can live with and after the market has already shown some confirmation. Once they are in the portfolio I want to leave them alone as long as the business keeps delivering.

2. Swings – shorter-term positions. I buy them with a clear idea of what I want to extract and I am willing to sell when that performance is reached or when the thesis stops working. Rivian is a good recent example of how I think about a swing.

In the past I mixed the two. I treated some positions like it had to become a 10Y holding or treated a core idea like a trade and sold it too early. Both mistakes cost me.

How to improve: Split your positions, mentally or practically. Have a different framework and conditions for each or own different assets with different strategies on different brokers to avoid mixing your liquidity.

3. No position is sacred

You will make mistakes. What matters is how fast you admit one. The faster the better, and the more impact on your returns. When something breaks, whichever data set used, accept it and act. No in-between, no “let’s see”.

Rule No. 1: Never lose money. Rule No. 2: Never forget Rule No. 1.

Warren Buffett

TFBTM - Intuitive Machines cost me a lot. While the space trade was raging I decided to buy the first pullback we’d had in a while. My condition to close the stock was the usual one: a weekly close below current support. When it happened, I was so sure it would bounce back I would just wait for it. My conditions were met but I didn’t execute. What should have been an 8% loss became a 24% loss.

FJR - New Fortress Energy was the clearest example. I knew deep down it was a mistake the first time I looked at the guy running it. Still, I stayed. I listened to the story. I watched the interviews. I wanted it to work. It went to zero. That one still stings because it was pure self-inflicted damage. I held it because I hoped it would come back. I had positions that were clearly broken but I kept telling myself the thesis was still intact.

How to improve: Have a plan before you buy. There are thousands of ways to make it in the markets but all of them can come with a plan. If you invest on fundamentals, know which data breaks the thesis. If you invest on price action, know which level is too low. There's always a line. Draw it, and respect it.

4. Treat your winners like gods

Here’s a contradiction. Losers aren’t sacred. But winners are. Hold them longer than feels comfortable and as long as they hold to your plans. The market isn’t made of math, it’s made of emotion. Winners usually run far past any model you built for them. Don’t clip its wings because a spreadsheet told you it reached “fair value”.

It never was my thinking that made the big money for me. It was always my sitting.

Jesse Livermore

TFBTM & FJR - Palantir left scars on both of us. We both bought in the teens and both sold at different levels because “multiples are stupid, it has to slow down”. It never did, and what could have been ~18x returns today became a few hundreds.

We caught the first wave and handed away the ocean because “we thought” it was too beautiful to be true, even if the thesis was still holding.

How to improve: The contradiction ends here. You have to treat your winners exactly like your losers. Same discipline, inverted. You bought with invalidation levels. Thesis probably got updated during the run. As long as they hold, do nothing.

Sitting is the skill.

5. Size is the real game

Stock picking is overrated. Your returns aren’t decided by which names you pick, they’re decided by how much you buy. A large position in a solid blue chip beats a tiny one in an obscure small cap. Your returns are defined by your ability to size up.

This is the one place math actually rules the market.

Position sizing is probably 70 to 80% of the equation. It’s not whether you’re right or wrong, it’s how much you make when you’re right and how much you lose when you’re wrong.

Stanley Druckenmiller

TFBTM - Palantir and Hims both come again, as, despites selling the first too early, I still made a great trade. But it didn’t count as I didn’t size it properly. I should have multiple times more money today not by better stock picking, but by better sizing.

This is what I changed the most over the last years. Back in early 2026, I made two large bets on great narrative and price action: Silicom and Soitec. Both combined now weight ~50% of my net worth and returned 100% and 150%. Those mattered, those increased my net worth significantly.

The stock picking was right, but the size made the difference.

FJR - What “sizing heavily” actually meant in my case: at one point Oscar Health was the only stock I owned. It was roughly 80% of my entire net worth.

That is extreme and I am not recommending anyone copy that level of concentration. It is simply what I did after fifteen years of trial and error.

The position was large enough that if the business worked it would meaningfully change my financial situation. That is the standard I now use for a core position.

How to improve: The fear of sizing usually comes from the fear of losing too much. This brings us back to planning. If you know your risk, you can size up. Not knowing your risk is exactly why people build small positions. And why returns remain low.

6. Less is more

Buying and selling constantly brings nothing but headaches.

Being everywhere, Trying to know everything, watch every headline dilutes your attention and your capital. The right number of trades is the fewest possible. The right number of stocks is the fewest possible.

Investing is counter intuitive and might be one of the only places where doing more doesn’t lead to better results or faster improvements. On the contrary.

The big money is not in the buying and selling, but in the waiting.

Charlie Munger

TFBTM - Soitec. Silicom. Nebius. Arista. Bitcoin. If I had focused on these five assets, which have always been some of my favorite assets on the market, My returns would have been much better. Add Palantir to it, and “much better” becomes “top 5%”.

I didn’t. I diluted my attention/liquidity to space. To healthcare. To SaaS. To sportswear even at a point. I kept following those, but had I done fewer transactions, less news and fewer stocks, I’d have done better.

How to improve: You only need one trade to set yourself for life. You certainly do not need 10,000. Tune out the feeds and the news outside of your framework. You’ll miss on some narratives. It doesn’t matter.

Just like you can’t be everywhere in life. You can’t be everywhere in the market either.

Final words

Over time the pattern became obvious. The largest chunk of our returns and net worth increase came from a handful of concentrated bets on businesses we understood deeply and sized properly. Everything else was a distraction.

These simple lessons took us years not only to understand but also to build an investing system around, as integrating them won’t be enough to improve your returns. You won’t need more work to set it up, but you will need work. On yourself.

Still, they could hold on a very small sheet of paper.

That single shift in mindset is worth more than any valuation framework.

The Few Bets That Matter and FJ Research are aligned on this. We both believe that a small number of high-conviction positions, sized properly and held with discipline, beats owning twenty interesting ideas. And both invest that way. Most investors spread themselves too thin and then wonder why the results stay average. Others find the right idea, size it well, and still sell it just as the compounding starts.

You only need one trade to set you for life. Don’t miss it.

FJ Research uses Revolut as my main account for investing, saving, and traveling across Europe, the US, and beyond. It’s the closest to a true all-in-one account.

Join me and over 80 million users here:

Join Revolut with RJF link

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

…

I’m here for the long term either way.

We’ve all been through this chain of thoughts, and we’ve all been stuck in a position we didn’t want in but felt like holding would be better than cutting. Because selling a loser makes you feel like one. It makes a mistake real, and you don’t make mistakes. You’re better than that. Actually, selling that stock would be the mistake, because it’ll go back up. You know it.

Except it probably never did. Bad became worse until you finally either chose to close your eyes forever or sold for much worse than it could have been.

This article was requested of me from a Founding Member after a note I wrote on a loser I cut early which continued falling, much lower. It’ll probably bounce back one day. But that doesn’t even matter.

That’s what we’ll talk about today. The importance of cutting losers early, how it can affect your portfolio, your psychology and your returns, and how to do it without entering this unhealthy emotional loop of feeling like a loser. But first, we’ll need to take a trip through concept-land, as cutting losers happens for a reason.

But I can give you the reason to do so now: closing losers early makes you a winner.

What Game Are You Playing?

This is the first question you have to answer, and there are no wrong answers. But there are some logical thought chains that cannot be avoided.

A stock picker’s objective is to outperform his benchmark within his timeframe, while sleeping tight at night. No one would ever pick a stock if it weren’t to increase his net worth. Right? No one. Surely.

Let’s define two words: outperformance and timeframe.

What is outperformance? The first mistake many make is to treat outperformance per position. Positions don’t matter. Net worth matters. If your net worth increases faster than it would have within indexes, you’re outperforming.

Now, timeframes. I said “within his defined timeframe”. Let me be straightforward. If yours is the next 30 years, if you invest in a position today for “the next decade” or things like this, you should stop right now and go buy an index. You have no clue what the next decade will be made of and cannot buy a position with the expectation to hold that long.

That’s bullshit. Great marketing, “The next Buffett” and all. But this isn’t stock picking.

If you do not stock pick to maximize returns in a reasonable foreseeable future, you do not stock pick, and you’d be better owning indexes; focus on capital preservation, not increase. And that’s fine, there’s no wrong in that. I’d love to have $10M and focus on capital preservation, but that’s not where I am today - not yet.

What is a reasonable foreseeable future? I don’t have an answer, which is why I do not have timeframes on my investments. I hold them as long as my thesis holds. Our world moves faster than ever, our economies too, geopolitics is messier… We aren’t living in the 50s anymore. We can’t invest as if we did.

My goal - and I believe every active stock picker’s goal, is rapid net worth increase. This doesn’t mean gambling, it means responsible investing constantly looking for the highest performer.

The market is the only place where anyone can make it, as long as he/she puts in the work and learns from his/her mistakes. It is the great equalizer. But you gotta play the right game if you want to make it.

This is about cutting losers. Don’t worry.

I'm a big Druck fan and see the market and my performance just like he does.

I compete against what I would call the opportunity set. And if there was a great opportunity set that year and I missed it, I’m disappointed in myself. Like if I’m up 20% and I think I should have been up 50%, I’m disappointed myself. If the opportunity set was basically to be up 10 or 15 and I’m up 20, I’m thrilled.

This is the game I am playing. Everyone doesn’t have to play the same, but if you do, there are some pre-requirements that makes cutting losers a must-do, for some very basic - and mathematical, reasons.

What The Market Actually Is

Millions of individuals exchanging opinions at a price.

That’s really all the market is, nothing else. Prices aren’t decided by opinion volume but by dollar volume. If 9,999 retails have an opinion and one fund has another, they can balance each other. This is an extremely important distinction as it changes what moves prices.

If liquidity is equivalent between bulls and bears, stocks range. That is what happens the majority of the time in the markets, lots of movement but no higher highs or lower lows. Just useless noise. When liquidity doesn’t balance, stocks go up or down depending on the winning side.

Our job is to catch the first and avoid the second.

This month was the release of the 30th anniversary pack of Pokémon TCG, with some exclusive cards.

Stick with me here.

Every physical store sold out in a day. Can’t find any box anymore, got to order it.

This, friends, is what happens in bull trends. In a bull trend, any acceptable price sees a surge in buying demand, which pushes the prices higher as bulls agree to pay more to be involved. Shares go to the bulls who buy the highest.

Official Pokémon's booster won’t be sold at higher price - although resellers will, the only solution to purchase is waiting for production. But if they were to auction the last packs, their price would rocket.

Why on earth is he talking about Pokémon?

Stocks never sell out.

This is when you make money, not by buying green dildos but by being positioned on the stocks with potential before the market sees that potential and bulls fight to buy it. The stocks with such potential are the stocks on which bulls constantly step up at key prices, because they accept to buy higher to be involved.

If bulls don’t step up, your stock isn’t a 30th anniversary celebration Pokémon box.
Its potential is limited, so why would you want to own it?

Some Truths You Won’t Like

There are four truths in the market I learned the hard way and I believe anyone serious about investing - the way I am, should accept as early as possible, because they’ll change your way of investing - and your returns.

You aren’t smart. Sorry, I know it’s mean of me. But it’s true…

If you were as smart as you thought, you wouldn’t be reading these lines and you wouldn’t buy stocks, you’d be working at Nvidia, Tesla or Lumentum and would be paid in stock options, already multi-millionaire or close to be.

But you aren’t. Your understanding of subjects and technologies is limited, your access to information is limited, your bandwidth is limited. You are limited - and of course I am too.

Relying solely on your opinion is the worst mistake you can make in investing.

Everything is relative. Your opinion - which is most likely wrong to start with and doesn’t matter, isn’t shared by everyone else.

I love chocolate. Everyone does, right? Wrong. Two friends of mine simply can’t eat it and looked disgusted the time they tried homemade cookies made by my mom. Shame on them.

We live in echo bubbles in the age of social and feel like everyone agrees with us because that is how algos work; they’re based on positive reinforcement. You like milk? Have more.

This is what I consider one of the most dangerous trend of our time, constant opinion reinforcing. But that’d be another subject entirely.

Stocks aren’t companies. They represent ownership of a company but aren’t priced for it. They’re the representation of millions of emotional reactions to news and expectations. A good company can have a terrible stock and can destroy your investment account.

Stop thinking that you have to buy good companies to make a killing, that’s wrong. You have to buy a good stock.

Good companies are a component of investing. Not the sole one. Plus, that relies on your understanding - which is limited, and “good” is relative - everything is. I consider Meta to be a good company. Do you? Because many refuse to invest in it due to its apps’ impact on the world.

Some of you will tell me “but even you look at good companies only”. I do, I’ll explain why just below.

My point here is that there is no correlation on your potential net worth increase and how “good” the company’s stock you bought is. Strategy can return 10x more than Google and yet I’m sure we’ll all agree on which is the best - Strategy of course.

Only price matters. You are here to make money, aren’t you? Then the sole judge of your success is your net worth increase, which happens when the stocks you own go up - I won’t consider shorting because that’s a stupid thing to do.

Nothing else matters.

As a stock picker, if you had a system that would work to buy stocks only based on price, you should use it. As humans, we need an illusion of control. We need to feel like it isn’t that simple, that we’re better, that our results came because of our sharp minds. That’s why we look at 10Qs, valuation, earnings, etc… In reality none of those matter to someone capable of generating returns based on price only.

Returns are the same for everyone, regardless on how they made them.

Why don’t I only buy on price? Because to concentrate in large positions, I need peace of mind. And I cannot achieve it by buying only a chart, I need to have facts backing it up. This doesn’t mean I am right, it means my psyche works that way, it means I also need that illusion of control. I need my brain to understand the narrative and the company so I can handle volatility and sleep at night.

It's a me thing, not a returns thing. I just know myself.

We’ll talk about closing losers now. But all of those concepts are inherent to why you have to close losers. I know it looks like I am off subject. Trust me, I am not.

The Math of Concentration

Something we’ve been over already. A 100% return on a 5% size is a 5% net worth increase. Youhou. I mean it’s good, but is it really what you’re looking for? Find a stock capable of doubling, and getting only a 5% net worth increase out of it?

Because you didn’t do 100% performance. You did 5%. Which will certainly be mitigated by the 19 other positions’ losses and be closer to 0%.

That sucks.

You have a certain amount available to you to invest. Whether four, five, six, ten digits… Doesn’t matter, you have a limit. You cannot buy 50 significant positions with the potential to meaningfully lift up your net worth.

If you have more than 10 positions on a 5-digit portfolio… You’re playing the screenshot game. Sorry, not sorry.

Real winners are rare, which means you have to size onto them. You won’t ever make significant money if you never size - and that is the goal, remember what a stock picker strives for? Rapid wealth increase.

You don’t have a choice to achieve this: you need concentration. Not all ins, but real size behind attractive risk/rewards that bulls are defending. Remember how to find those? When strong levels are defended because bulls are fighting to buy at the best possible price but also before other bulls, not to miss the opportunity.

A $10,000 portfolio will be happy with a $3,000 gain, right? That’s a 30% position into a stock with a potential of 100%. You could have three of them, two successes would generate $6,000 returns. Now we’re talking.

But in every portfolio, there are winners and losers, and rarely a 66% win rate.

I mean. I know you’re the best stock picker on earth. I am too. We rarely miss, don’t we? Because we never sell. We just hold and wait for the bounce because we’re right, the company we bought is a great company and the market will realize it. Soon.

Wrong.

The problem is that concentration works both ways. It increases gains, but also losses. A 25% loss on a 30% position would be -$750. Meaningful for a $10,000 portfolio which became a $9,250 portfolio. Generating meaningful returns got harder because your positions got smaller.

You see this table a lot on social media. Print it in your brain. Concentration raises risk. And losing capital makes winning so much harder.

Let Me Tie It All Together

Because I didn’t lie to you. Everything is related.

You want to grow your net worth fast. That means concentration. But the market is a place where millions of opinions are weighted in dollars. The big fish move it. They know more and if not, their sheer size will push prices wherever they want. Your convictions are unreliable by default.

You are playing a losing game. Forced to play big on unreliable assets.

A stock falling isn’t noise, it’s the bigger players telling you you’re wrong. Maybe they know something you don’t. Maybe they’re just heavier. Doesn’t matter. Price is them voting against you, and they always win.

So you cut. Fast. It keeps the math on track - so one wrong bet doesn’t break your recovery, and it protects you from yourself - from the ego that insists you’re right when the whole market says otherwise.

Concentration is what makes you rich. Cutting losers is what lets you concentrate without blowing up, against a market smarter and heavier than you’ll ever be.

And you know what? You’re better off in another stock either way.

Who Cares If It Bounces Back?

This is screenshot mentality, another gift from social media.

How to avoid the frustration of missing out and ego spikes? By realizing that it doesn’t matter, first, and that you still can do something about it, second.

One position is just that: one position. The amount of opportunities in the market is mind-blowing; if you’ve followed me for some time, you see that I can share a handful of them per month, whether new or accumulations. You’ll never be short of opportunities and stocks with 100% potential. Never. Even during bear markets, these exist.

Looking back at a loser after selling it ignores the most important part of investing: psychology. A bounce doesn’t mean you’d have held during that time.

AsteraLabs was one for me. I was dead right on the thesis. I sold it as bulls didn’t step up ~$130 . The rest is history.

I take comfort in knowing that Jamie caught it. Hi Jamie.

You know how many AsteraLabs I had since I write this Substack? You know how many didn’t bounce back? I don’t. I don’t care. My portfolio is largely outperforming and I am extremely satisfied with my net worth increase, which means I’m doing things right, even if AsteraLabs pumped without me.

I still had more losers that kept losing than losers that bounced back. And we’re not talking about 40/60, but 80/20. Imagine the net worth decrease if I had held them all. I also couldn’t have bought my current winners with large positions. My performance would be crap…

Second, what happened to that liquidity is what really matters, because you are here for net worth increase, right? Or for screenshots? Be sure.

Selling one losers means buying or accumulating another position. What happened to it? Did it win even more? Were you able to sleep at night better? Did you stay in cash waiting for another opportunity?

Investing isn’t about individual decisions. It is cumulative. A chain of decisions.

Maybe you let one winner go. But if selling it was at the price of respecting a system that saved you more times than it hurt you, then you made the right call, and you should hope to make it again. If not, one day, you won’t cut one of your losers. And it’ll hurt really bad.

If you constantly cut future winners, your system might need some review. But be sure to have it quantified before changing anything. Survivor bias is real, so let’s not make the mistake of considering one rare occurrence to be the norm.

That being said. Selling winners happens. But there is a silver lining here, something we can do about it.

We. Can. Buy. Again.

Another psychological barrier in investing. Apparently, once we cut a loser - and we accepted it was a loser which is already hard, we can’t tango with it again. This is bullshit. You can go back ten times with your high school sweetheart but you’ll tell me you can’t buy a stock twice? Come on.

If that stock you sold because it behaved poorly bounces back and gives you one of those opportunities, just take it! I’d have loved to buy AsteraLabs again and was actively looking to do so. I was never given the chance, so be it. My cash was in Soitec by then - rightfully so.

Two stocks gave me a second chance after I cut them: UiPath and Tempus. And my portfolio’s happy I let my ego aside and bought them back.

Selling isn’t defeat. It’s responsibility. You’re just a dude respecting the market and preserving your capital to fight the next battle. You aren’t giving up, you don’t suck, you aren’t bad.

You are responsibly allocating capital.

So, How To Do It?

There’s no easy answer.

This post is meant to give you my thought process, starting from how I invest and why because cutting losers is inherently connected to my system and objectives. It’d be impossible for me to invest the way I do without cutting losers fast.

The psychology behind my loser cutting is printed within my investing system. I’d never overperform without doing so as my liquidity would be tight up in sluggish or underperforming stocks.

If you invest differently. So be it. I personally consider that cutting losers is a must-do regardless because capital preservation should remain every stock picker’s #1 rule, but you are entitled to your ways.

Everything is relative, eh?

Know what you want. Know how you want to do it. This is how you psychologically accept to cut your losers fast, by knowing what you’re doing. Because eventually, you simply won’t have any other choice but to do so.

Practically, I can give some tips on how to minimize your chances to fall into this psychological “let’s see later” trap.

1. Make a plan before you buy. You should always know why, when or how you would let your position go. There is always a place where bulls should step up within a healthy trend. Always.

2. Make the loss concrete. Look at the potential dollar value loss in your portfolio before you buy a position. How hard will you have to work to refund your loss?

3. Don’t second-guess yourself. The plan was made in an emotion-less situation. If the conditions were met, execute. Close, move on. Order 66 style. You’ll sell future winners, this isn’t even a question, it’ll happen. But you’ll sell more large losers and that’s enough to justify missing out once.

4. Selling means reallocating. There are always opportunities and the best ones are usually the ones you already own, as it means they performed very well - otherwise they’d be cut. Nothing in investing is about single decisions, it all is a chain of decisions.

5. Keep an eye on strong convictions. If you were right, you will be given another chance. This isn’t goodbye, this is see you later.

Closing Thoughts

I’ve been right on so many stocks over the last two years - notably Palantir Nebius and Hims, I should have multiple times more money today. This isn’t me being cocky, this is pretty factual. I don’t, because of two factors.

1. I never sized aggressively enough.

2. I cut my winners too early.

I thought hard at why and how I could fix it.

• The first was because I was scared to lose too much, so I middle sized everything. Cutting losers with controlled risks allowed me to use size I’d never have dared use because I knew my maximum potential loss before hand.

• The second was because I always wanted to buy something else. Cutting losers allows me to have a pocket of liquidity always ready. If I don’t have, it’s because I only own great winners and I now think twice before rotating, because I became aware of it.

In investing, you only need to be right once or twice per year. But that is true only if you size your trades properly. Being right a handful of times that way will set you for life, at one condition: that your losers don’t drag your winners down.

And that can be done only if you cut them fast.

Important Note: The tone of this write-up was voluntarily harsher than usual. I usually try to keep things very smooth and professional but with this kind of subject, I believe harsh words are more impactful and therefore more helpful than sweets. Do not take anything personally, I in no case mean to be hurtful or violent.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

So sometimes, the market is just being stupid. And when it is, it hands us one of the best companies in the world at a discount. We’ve seen it, and bought it, with Meta, Amazon, Arista, Tenable and many other names in the portfolio this year, pushing returns to 182% since inception in March 2025 and 98.7% YTD, transparently.

And it is happening again, with one of the biggest AI names in the market.

A company growing its AI revenue over 200%, guiding for triple digit growth, sitting on more demand than it can ship with visibility running not for the next quarter, but years. A stock the market sold double digits because of a guidance that missed for less than a percentage point.

That’s not a fundamental problem. It’s a market problem. And market problems are where the best risk/reward is born.

Below, I break down this company, what it does, why the demand is real and how to know it’s not just management talking, what the numbers could look like, why the market isn’t buying the story just yet, what worries it, and exactly how I’d play the setup the market just handed us.

Read more

We’re about five months into the premium version of my Substack, and I want to start by saying thank you, again. I didn’t expect these results from a small, slow-growing subscription like mine. Your trust and commitment genuinely mean a lot.

I’ve been working on improving everything lately, content structure, formats, write-ups quality, and how the chat is used; ithas become a real place for questions, ideas, and live alerts. If you’re not in it, you’re missing out on a big part of your subscription.

I hope you guys see the improvements and appreciate them. This goes on the direction of this note, which has two purposes: asking for your feedback, and show you something new I’ve built, which many of you asked.

First things first, feedback

I’d love to hear about your experience. What works, what doesn’t, what you like and dislike, what you’d change... I won’t act on everything, but Founding Members know I listen and can adjust when it makes sense.

To make it easy, a few questions, feel free to answer any that speak to you:

1. What’s the most valuable service you get from my Substack? (And the least?)

2. Which is your favorite and least favorite format, and why?

3. Are the investment calls and plans clear and actionable enough?

4. Are the write-ups well structured and informative enough?

5. What would you want more of? Earnings reviews, new positions, positions reviews or update, macro, something else?..

6. What’s missing? What’s unnecessary? What’s too much of? What’s not enough?

7. If you’re a free subscriber: what holds you back from subscribing?

I’ll give away a few free months to those who take the time to give real feedback. Your input genuinely shapes where this Substack goes, its formats and services.

Second, my investing system. Undressed.

My investing system that generated 182.3% since March 2025, compared to 38.6% for the S&P 500, and up 100.2% YTD. Now transparently accessible.

I’ve entirely rebuilt my tracking sheet. It now shows every transaction, tied to the alerts shared in the chat and going back to early 2025. A complete, day-by-day view of the Substack portfolio.

Let me explain what it is, and how it works.

The SS Book is a live representation of my investing system. My investing is simple: build large positions at great risk/reward, average up when they behave, organically concentrate into winners, cut losers fast. That’s what I do, and the SS Book shows it running in real time.

That book’s performance is the representation of the value-added of my Substack. All positions management shared, transparently, whether you took the trades or not.

My own book is liquidity-constrained - we all are. I can’t buy every good setup. I’d rather hold Silicom than trim for another name. But all of you do not have a 30% position in the name from the $20s, which means your choices will be different.

The SS Book isn’t constrained that way. It captures every call and alert I share, owned or not; the truest picture of my system’s stock-picking, not just what I happened to have cash for. And as our books are all different, this is the best value for you as it includes many more great opportunities than only the one I take, better than if I were to only share my own book and transactions. But know that all my personal purchases are logged into the SS book, as I follow the exact same rules.

Fixed rules, a robotic investing, stripped of emotions and bias, but still liquidity constrained. Here how the SS book works.

1. Positions have a standard size ($10k or $20k for blue chips at exceptional risk/reward).

2. W50 accumulations once a month for $10k; aggressive accumulations twice a month maximum for $5k.

3. Trims of 30% on extension with proceeds reinjected on the next purchase - not a fixed inflow post-trim.

4. Returns are dollar-weighted on net capital, because timing and sizing are part of the system.

Now you can see exactly how I manage my positions; every trade, every date, every alert behind it. In investing, the value doesn’t come from opinions, contrary to what many think. It comes from execution, buying at the right place with the right size, something no one shows or talks about.

The dashboard still requires a bit of design, but I wanted to ship the data as soon as possible. Paid and Founding Members will have access to it through different links as the FM sheet also includes my portfolio weighting, to see which assets I concentrate my liquidity in, and my buying targets, updated weekly. Nothing else changes.

You’ll find the different links below.

This is meant to give all paid subs a concrete view of my system and its value added, with concrete numbers and position management, allowing them to review all trades and anticipate coming ones.

I hope you’ll find value in this new data set.

Don’t forget to give me your feedback on my Substack’s and this new service. It’s all built for you, to share as much value as possible and improve our returns, together.

The grind continues.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

It was skeptical on AI and earnings for long. Most companies beat and AI is delivering above expectations, we’ve been over this. It’s now skeptical on AI itself, not that it won’t work - which used to be fear number one, but that it would work too well and become dangerous, which is the latest narrative. And it’s skeptical on rates, worried the Fed hikes this week.

None of it changes what I’m doing. Here’s why and what I’m actually buying.

AGI is Here, The World is Doomed

This is coming from the frontier CEOs themselves, with random posts on X coupled with some interviews where they talk about how advanced models are and how important it is to be more reasonable.

What. A. Load. Of. Crap.

I don’t mean to say that they are wrong and there are no dangers with AI. There are, just like with any new technology. But that they would become reasonable now just for the sake of humanity is certainly not true. This is excellent PR, and they certainly have many reasons to do this - some I can speculate on. But certainly not the ones they claim.

As for this kind of dramatic claims… There’d be lots to say about them as well, but while Dario and Sam have a great PR, this was terrible PR as the current view of AI by populations isn’t really positive lately between electrical consumption and job cuts.

Hopefully Jensen and Karp are here to bring some critical thinking to the table as this kind of claims bring nothing - whether they are true or not. So let’s do some ourselves.

First, research never stops. It never has. The Manhattan Project didn’t slow down for danger, it sped up, because otherwise someone else would have made nuclear fission work first. China won’t slow down. There’s no world where everyone agrees to halt AI, not before something bad happens. You build it before your rival does. That’s how everything has always worked, even if we don’t like it…

If we didn't have adversaries, I would be very in favor of pausing this technology completely, but we do.

A. Karp, Palantir CEO, CNBC interview

Second, these CEOs spent years hyping AI. Their approval’s tanking and they need political cover to keep building. Sounding careful is the move after telling the world that AI will cut 50% of blue collar jobs. They haven’t decided to slow down, they just try to look like it.

Third, it doesn’t even matter. The money isn’t in more intelligence anymore. Current models are enough to unlock trillions of value if used properly. This is the next step, optimizing them, not training more frontier models. Do you really believe they’d slow down monetization? Never. They need cash, and they’ll find it whatever the cost, they just know they don’t need to go further in training to generate it.

Quiet research, PR stunts, pivot to optimization. Not a slowdown, a reprioritization at best with some good optics and added soft power to lobby. Cherry on top. If the market is stupid enough to read this as “AI spending is over and AI services are done”, we’ll be given so very juicy gifts.

Because the truth is that use-case optimized models will flourish and generate an accelerating demand for compute which requires more hardware and data centers to be met.

Nothing changed.

Let’s see how the market reacts to those positions because at the end of the day, we only make money if the market agrees with us. Being bullish in a downtrend won’t make returns. But understanding the mechanics behind will allow us to take advantage of a market’s overreaction.

It might not happen. But if it does, I’ll be buying.

The Fed Hikes

The market’s now pricing a hike this week after a hot MoM core CPI, Warsh removing guidance and some members publicly stating they’d advocate for a hike.

The Fed’s mandate is to react, not anticipate. One hot MoM print after months of “good” data and no yearly acceleration doesn’t seem enough to hike. The market will price the worst every time, which puts us in a pretty comfortable spot.

Bad news will confirm pricing. We’re there already. Good news on the contrary would be completely unexpected. The market isn’t ready for no hike.

Here’s What I’m Doing

The market is giving us good names at good prices. And I continue to do the boring thing: buy strong setups on support, ahead of data everyone’s dreading. Because those are the best risk reward on the market, winning big if I were right and losing small otherwise.

Below: my new positions and three other actionable setups. Healthcare, hardware and software. Strong sectors, great companies and strong price action.

Explosive setups.

Read more

I bought the stock on the previous breakout, a few weeks ago. The logic was simple; SaaS were going to become a big deal once AI-enhanced and we started to see the market treat them as such with the IGV breaking out early May. Many leaders broke out and shot up, becoming too expensive before any fundamental confirmations. But UiPath was just starting its move, a bit delayed.

My first position got stopped out rapidly on Claude news. I took the trade a second time as the setup got even better. The position is now up ~19% after this 16% drop. An extremely good performance in a month.

We now need to review the quarter as it has some impacts on the thesis, and what to do with that position.

Saaspocalypse Is Cancelled

AI demand isn’t a question anymore. We know it’s here, we know it’s growing, we know it’s massive. We also know it isn’t showing up everywhere at the same speed.

We have clear winners - Palantir and cybersecurity, who keep winning. Their numbers continue to accelerate. But many SaaS are still either in the earlier innings with some numbers that start to ramp up - like Zeta. Or showing healthy core business which invalidated the bear case, but no AI acceleration just yet – ServiceNow as an example.

Overall, we know where the SaaS market is going, in time. (up).

So when UiPath falls 16%, it isn’t because results were bad but because in a market, everyone looks for performance and performance comes with acceleration. This quarter didn’t show acceleration - while others did, so the market let it go.

The question was never “can UiPath do well in an AI world?” We know the answer: yes. The question is: can UiPath do above average? Because that’s what we need to have a performing stock.

This quarter, the market voted no.

But the market votes wrong all the time. Many stocks deliver above average returns without above average results, because valuation and momentum are also part of the game. Not just fundamentals.

Let’s look at the data itself, the potential of both the company and the stock, and what I’m doing with my position.

Read more

Strong trends usually climb in steps: optimism builds, expectations rise, stocks rise until profit-taking starts and the market corrects its own optimism. From there either the optimism was misplaced and bulls will let go, trends break, and that’s it. Either it wasn’t, and a new leg will start once bulls take over the tape and start buying strong supports again. Between uptrends and downtrends - which is the case 90% of the time, the market is choppy. Short term volatility up and down, not breaking any trends, just uselessly moving.

May to July was the market breathing and correcting its own optimism. it got carried away, traders took profits and funds went to liquidate Leopold. Since then, the market pushed on optimism after the liquidation news, but what we are seeing isn’t the return of an uptrend, but choppiness between strong supports and resistances.

As expected, ranges are forming.

Bulls and bears are fighting on different data at the moment, with valid reasons on both sides.

On one side, inflation remains “too high” and Kevin Warsh has left the door open to rate hikes depending on the data while the situation in the Strait of Hormuz remains uncertain for energy prices and inflation.

On the other side, AI earnings are extremely strong. Companies are raising guidance across the entire supply chain, from hardware to software, reporting more demand and proving that customers are getting real value from AI, etc...

Macro environment is uncertain, as always. Micro is improving. That’s the battle the market is trying to resolve right now: macro versus micro.

I’m betting on micro.

The correction brought high-quality companies back to levels where the risk/reward is more attractive than they were a few months ago. This is what I wait for: better fundamentals, lower prices, great risk rewards with a clear invalidation.

Optical names are a good example. Many of them corrected sharply from their highs while demand for optical connectivity continues to accelerate. I’ve already shared a few that have delivered good returns including AAOI at $91, ANET at $140 or $SOI ~€90. I’ll share another one today.

Photonics used to be what everyone wanted to buy. Now is the time to do so. An upstream company, directly exposed to demand for any kind of AI-optics, now sitting at the support level I’ve been waiting for, for months.

Read more

This quote from Vladimir Lenin had nothing to do with the markets. It was about how fast revolutions can transform a country, an economy, or a society. But it fits the market quite perfectly.

The last two months were some of those weeks. A massive drawdown, a rapid recovery, dozens of earnings, countless setups, failed supports, and a portfolio going from -30% to all time highs in a matter of days.

These periods also come with more transactions than usual: letting weak names go, accumulating winners, and buying strong setups, all while positioning the portfolio for the next phase.

That creates a different challenge: not finding more things to do and opportunities to buy, but knowing when to do nothing.

Before starting, if you aren’t part of the chat, you’re missing out on daily updates, market commentary, and transaction alerts. This is where I post all important day-to-day information and transactions, with value added for all subscribers.

Consider joining!

The Biggest Questions Were Answered

Over the last three years, every quarter had a concrete bear thesis.

The first was about hardware. Could GPUs allow parallel computing for intelligence, not just graphics? Nvidia’s Hopper delivered, and the later generations brought leaps of improvement, very rapidly.

The second was CapEx/ROI concerns. Will spending lead to cash generation? This was worsened by debt later. Without accelerating cash generation, companies could end up insolvent at worst, slowing down their future growth at best. It took quarters to have answers. The previous earnings season gave clues. The latest one sealed the deal.

Lastly, the most important of all. Can AI generate final user value? If not, nothing else would matter, as no economically viable service means everything was done for nuts. This is the current question.

The question left isn’t “does compute have ROI?” but “does AI have ROI?”. This won’t be answered by hyperscalers, but by SaaS, as they are the ones consuming tokens to deliver services. Considering the amount of tokens they’re using, I’d imagine they see ROI from it but we are yet to see the real numbers.

But this season gave us an answer.

Salesforce, CrowdStrike, Rubrik, ElasticSearch, Sentinel… They all confirmed one or more of the following trends: revenue growth acceleration, margin expansion, and raised guidance. And all cite AI as the reason for those fundamental improvements.

A few weeks ago, ServiceNow proved that AI won’t disrupt their business, without proving that it’ll enhance it. The largest bear-case fears were reduced while most of SaaS valuation got crushed, and the market started buying those names again as the risk/reward became attractive. It’s now even more so as other companies confirmed this, but also added clear tailwind from AI.

The risk/reward became even better and the market should start to anticipate success. It already did for a few of them.

The Most Important Take Away

Simple.

If SaaS generate ROI from AI, AI spending will accelerate.

If the last link of the chain generates value, the entire chain becomes valuable as every dollar spent on AI - whether on hardware, compute or else, ends up generating value for one and ROI for another.

Palantir is a great example - the only one until now. Their software provides value to their customers, so those increase their spending and new ones reach out. Any kind of customers, for many kinds of usage within different industries. Other SaaS took a bit longer but are getting there now, and we’re yet to see consumer SaaS leverage AI as they should - companies like Airbnb or Duolingo. But they will.

AI SaaS demand increases → Compute demand increases → Infra demand increases.

The market expects continuation. I expect acceleration.

We Already See It, Even if Some Refuse To

Look at Nvidia and Marvell earnings.

The latest narrative was that Nvidia would crumble as hyperscalers turn themselves towards customed chips. That’s not what the numbers show. They show another leg of revenue acceleration, raised guidance, and stable margins.

Why? Because hyperscalers aren’t the only actors anymore.

There is sovereign AI, there are regional AIs, there are neoclouds, there are AI startups at enterprises [...] which represents about half of our business, and that is growing 100% a year.

Yes, Nvidia’s losing market share. But that market increased manifold in the last two years, which means a smaller portion of a much larger cake, which they are expanding into with new products - networking, cloud, software, acquisitions & investments…

There is no slowdown.

Marvell shows the other side of the piece as customized hardware accelerates - as expected since months. Record earnings and increased guidance, which reinforces my convictions in most of my current positions, which I accumulated over the last weeks as shared with Founding Members in the chat.

On the optical DSP side, 800 G demand remains strong, while our 1.6 T business is ramping rapidly, a trend we expect to accelerate further in FY 2028. Within scale-out switching, our business remains on track to more than double this year, driven by a strong ramp in our 51.2 T products across a broadening array of customers.

However, the significantly higher bandwidth density required by scale-up networks is best served by bringing optics much closer to XPUs and switches. While the transition in scale-up networks from copper to optics is expected to take several years, with both technologies coexisting, customers are aggressively planning scale-up optics deployments starting as early as next year.

We expect this growth to be driven by both our XPU and XPU-attached products. In XPUs, we continue to make strong progress across current and next generation programs at multiple hyperscalers. In XPU attach, we are benefiting from increasing demand for both CXL and Custom NICs.

No slowdown. And now that we know compute adds value at the end of the chain… Why would it? I stick with my opinion shared a month ago, with stronger convictions after this month’s earnings.

We are going to see a bubble, but this isn’t it. A bubble is built on endless optimism, and blowing up a leveraged fund by selling amazing earnings showing clear ROI from AI, revenue acceleration, and margin expansion isn’t really what endless optimism looks like.

As for how to position yourself for this acceleration, the good side is that the market is giving you more opportunities as hardware isn’t the only benefactor of AI spending anymore. I’d expect it to remain the leader even if most are priced to perfection, some aren’t, and perfection’s temporary as analysts suck at predicting. They’ll continue to beat and raise as the cycle accelerates.

Software continues to show signs of reversal in both price action and fundamentals. Cybersecurity were the golden goose, but a few are only starting their uptrends now. Data providers and optimizers are also starting to show demand, final-product SaaS too… Liquidity will flow to them, and not being part of it would be a mistake.

Crypto continues to show signs of revival, even though it doesn’t fit into any of the current narratives. Liquidity goes where it chooses to go.

Those are the strongest sectors today. I’m positioned for the three of them, with very strong foundations for the months to come as the cycle continues.

From here, the best thing to do for me is nothing.

More Activity Means Less Returns

Downtrends are hard to manage. But uptrends are even harder.

When you found your winners, the best thing to do is just letting them run. Your job as a stock picker isn’t to constantly manage positions or look at every green candle, but to find the few stocks that will outperform the market, build large positions, and let them run as long as they behave, while cutting the bad seed fast.

This is less complicated than it looks; the hardest thing to do is leave your ego and pre-conceived ideas of what “investing” is behind. The market is what it is, not what you want it to be. Accept some facts; it’ll help you let money compound.

1. Valuation doesn’t matter. It is a personal opinion on what “should be” that you try to apply to a voting machine based on sentiment and billions of daily votes. You eventually need to accept that you don’t know better than the market… Even if you did, you’d still lose. An even harder and broader truth is that none of your personal opinions matter… I know, it’s a hard pill to swallow. The sooner you accept it, the better.

2. Price action matters. It doesn’t make you less of an investor to learn and use it. In a game based on votes, ignoring the visual representation of the voters is pure ego. I’m not talking about Ichimoku and other shenanigans. I’m talking about price and volume, about the factual data of what people buy, where and how. It’s much better data to use than valuation, as it doesn’t rely on a personal opinion.

3. Weakness compounds. Just like strength. I’ll give you a concrete examples below.

4. Sizing is key. What matters to me is to increase my wealth, not invest $1,000 on a speculative asset that could 10x. I’d end up with $10,000. Cool. Not life-changing money. You’ll make money on large positions. Not small ones.

5. Less is more. Large positions mean fewer positions. And that’s ok. You don’t need to be involved in everything. You don’t need to compare yourself to others. You just need to own large positions in stocks that beat the market. That’s how you increase your net worth, which ultimately is the only thing that matters.

Weakness Compounds

One of my Founding Members had a comment on Tempus a few days ago, stating that the stock was up ~40% since we cut it for weak price action. Factual. But there’s more to it than just a performance number and while I answered, I thought it would be a great case to look at in an article.

I sold Tempus on July 21st for a 7% loss as it broke below its W50. I’d be up ~22% today if I had held my position, but I also would have gone through a 30% drawdown with a stock breaking all of its uptrend supports.

Holding this stock here is “I think I’m right even if the market tells me I’m wrong and that position negatively impacts my net worth by 10%”. This is a hard psychological situation to be in, and while many could think “let’s see, whatever”, I’d argue this is real gambling. Capital preservation is too important to behave that way.

Then came positive healthcare news and Moderna’s positive Phase three skin-cancer results, giving the sector a major boost. Tempus jumped ~70% from its bottom, reclaimed its trendlines, and broke out to new highs.

A piece of news from another company triggered optimism. Great, but I could have gone to the casino and put red at this point. Still. Factually, Tempus is up 22% since my original purchase price.

But focusing on that number misses the entire forest behind one small tree.

What Happened To My Liquidity?

Selling a stock means that cash goes to another one. For me, it was AAOI at $91, which pushed 60% before I trimmed and held the rest - still do. The overall returns on that liquidity were similar, but I removed the stress of a 30% drawdown from a position that had lost its trend and had no clear catalyst.

Who knows if I would have held? And not sold the bottom, and missed out as well? And there is a second important parameter to take into account.

What If It Didn’t Bounce?

Survivor bias says that we have a tendency to look at the cases that worked out, while ignoring the ones that didn’t. Tempus bounced back. It could also have not. I cut loose many positions this year: Fluence, Enphase, SolarEdge, Cerebras, RocketLab… All of them are much lower today than when I cut them. They might bounce back. They might not. And if they do, what really matters is what will be done with this liquidity until the bounce, not that they bounce.

But there’s another statement I can make, which doesn’t look at one position, but at my total portfolio. I had held all broken trends, I wouldn’t be at all-time highs today.

Tempus was a loser. It now has much better price action and a better profile today, while its sector continues to be strong with many tailwinds. It became a much better candidate for a large position, maybe later.

You don’t make money by being right about everything. You make money by keeping losses small, putting size behind what works, following trends, and giving time to your winners to compound. Right now, my job is simple.

Hold winners. Cut weakness. And don’t confuse inactivity with doing nothing.
Sometimes, doing nothing is the system working exactly as intended

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

We should be moving from a downtrend to a fight between sellers and buyers - a range. I struggle to see sellers being stronger now considering the environment, but we shouldn’t rule it out. And I struggle to see buyers overwhelming sellers right away, but that is more likely than the contrary.

My expectation remains a range, as it has for weeks. But I am now more confident about the outcome direction of that range - higher, and its length - few weeks.

I didn’t expect the market to go straight up.

Even though that was a possibility, from a risk management perspective I preferred to let the market prove itself before getting aggressively involved. A healthy uptrend remains my base case today. Not yet confirmed, but increasing confidence.

The important part is that we are now getting what I wanted to see: a retest.

Most of the companies I have been covering broke out, proved their strength and are now coming back to test support. Volatility is violent but the trends remain largely intact. We now need buyers to step up on those supports and confirm my opinion of an healthy uptrend starting by doing so.

The S&P is the clearest example.

We had almost two months of range while funds worked at liquidating Leopold, followed by a breakout to new all-time highs once it happened.

After a move like that, I expected traders and holders to take profits or simply get out.

At this point, I see two major liquidity pressure points left:

• Citadel and traders offloading some positions on pumps, which could put some pressure on these names - and the global growth market.

• Retails and funds now buying back great fundamentals on great earnings and a healthy environment - except for Iran, cleared of most leverage and risk now that valuations have taken a large dump.

I’d expect the second to be the strongest, but I’d also expect the first to give buyers a a hard time short term.

That isn't a bearish chart. It’s what you want to see after a strong move higher.

The Nasdaq is weaker because it was more exposed to the liquidation, but it is still ranging between its previous high and its W21. Undecided, yes. Bearish, no.

If the market is indeed going higher, this is close to the best risk/reward.

You don’t chase a stock 30% above its support. You buy names that have proved they can move higher at levels where buyers should step in. We already have winners from the previous moves, some held for months like Soitec and Silicom, others bought during the latest bounce like AAOI, Harmonic and UiPath.

And earnings continue to confirm what matters: AI spending is translating into real demand, revenue and returns - proof 1, proof 2 and proof 3.

We also got another confirmation this week. Anthropic and OpenAI continue to grow revenues and ARR at double digit QoQ, with no evidence of the “slowdown”, circular financing or bad economics that have been used to justify a bear case for years now.

I have no doubts that we’ll have a bubble and that it’s forming.
But I believe the best is yet to come, not behind.

So what am I buying?

The market is giving us what I’ve been waiting for: retests.

It needed to digest the move. Strong names needed to come back to support. Traders needed to take profits and short. New buyers needed an opportunity to enter without chasing.

That’s healthy.

Some of today’s moves look violent, but I’d much rather have this than another straight-line move higher.

This is where you respect the system, no matter how hard it feels. That’s why you have one, to avoid making decisions on emotions and bias, and to buy based on concrete data and setups.

You don’t need to know whether the S&P goes up or down tomorrow. You need to know which setup offer the best risk/reward and how much you risk by taking it. If it fails, it’s alright. Can’t win them all. If it works, you’ll look back at this moment as one that shaped your year to date performance.

So let’s go through the best setups.

Read more

I didn’t call the bottom. I didn’t increase margin or use options. I simply went back to following my system: cutting weakness and concentrating on the strongest setups, increasing exposure as winners prove themselves.

The recovery came fast because the market gave us exactly what we were looking for: strong trends, clear support levels and earnings that confirmed the thesis. The key wasn’t knowing when the bottom was. It was being willing to act when the market proved us right, and cut losses when things got hard.

That is what I’ll review here: what happened during those two weeks, which positions drove the recovery, which decisions worked, which ones didn’t. And more importantly, I want to show why the system mattered more than any individual stock.

The warnings were here,

June was an obvious top in retrospective with two of the most important indicators flashing red: euphoria on social medias and extended stocks/indexes. This is when I wrote this article titled “The Breather is Coming”. Telling.

Not like I didn’t see it coming.
And yet…

I broke my own system

The drawdown wasn’t caused by one bad investment. It came from a combination of mistakes.

First, I overtraded. I had margin on and kept trying to buy setups while the market was telling me nothing would work. I’d understand why weeks later, but the reason didn’t matter. The signal did. When good setups consistently fail, the market is telling you something. Environment comes before individual stocks.

Second, I didn’t cut some losses quickly enough. The space sector corrected before the rest and I thought those names would bounce earlier. They did, but a month and a half later. That wasn’t useful. I had to take the trades; I didn’t have to hold them.

Intuitive Machines is the clearest example. I bought ~$20 on a W50 retest, with a strong narrative, a strong sector and a clear technical setup. I don’t regret taking that trade. It met all my criteria.

What was wrong was waiting until $14.70 to sell it, turning what should have been a ~11% loss following my system into a ~27% loss. Buyers didn’t step up where they should have - on the wick below the W50, so I should have listened. If buying zones aren’t bought, stocks go lower.

Third, margin. This one is more personal, but making those two mistakes while using margin obviously multiplied their impact…

The important point is that none of those mistakes required better stock picking to fix performance. They required better discipline.

Then the market changed

The drawdown had no bounces. Constant selling. Normal downtrends usually come with bounces and their absence was the most important information. Something happening. We just didn’t know what - we now do.

It marked the bottom as the selling pressure disappeared. I couldn’t be sure it was the case then, but for sure the risk reward had shifted. During inorganic selling, when the objective is reached, the selling pressure disappears while the strongest names were back at key supports. Back then, I wrote,

Talking about market mechanisms, chances are we’ve seen the bottom. This doesn’t mean we’re back in an uptrend, but means we’re good to buy again. And I believe the next few days will be key to the rest of our year’s performance.

Most of the selling pressure is behind us. Leopold has nothing left to sell, Citadel is now holding the chips and might - should, sell some, but not everything considering Leopold’s portfolio quality and latest earnings. The portfolio might be a bit outdated, as he had to sell some holdings, but its quality is undeniable.

That distinction matters. I didn’t need confirmation that the bottom was in. I needed confirmation that it was becoming worthwhile to buy again. I already had realized my mistakes and fixed them. I had cut losers, reduced margin and freed liquidity. I was then only looking for the best setup.

That was the turning point.

Concentrating on the winners

Nebius had already given us a clean W50 retest. Akamai was back at its W50. Soitec and Silicom were holding their trends. All of them were mentioned with subs. But one setup stood above the rest: Applied Optoelectronics.

The stock was retesting its W50 within a clear uptrend, with clear tailwinds and strong fundamentals/sector. I bought heavily.

Two weeks later, AAOI is up ~67% from that purchase.

This is where position sizing becomes more important than finding the best stcok.

AAOI wasn’t necessarily fundamentally superior to every other position in the book. But it was offering the better setup at that moment. I had a clear and very positive risk reward.

That’s the kind of situation where I go big.

The risk was clearly defined, not making the same mistake than Intuitive Machines. If buyers didn’t step up on this wick ~$76, I’d be out. A ~10% risk for a multiples in potential gains, assuming the trend holds.

That is very different from blindly buying a stock because I like the company.

You don’t need every trade to work

This is also why I don’t care about having a perfect hit rate.

The same week gave us AAOI HLIT PATH AXTI BE ON and a few more opportunities which I shared with subs. Some worked really well, some never gave the entry we wanted. Others, like Intuitive Machines, struggled to get their W50 back.

That’s alright. A ~10% loss is manageable if your winners can compound at multiple of that.

This is also why I don’t spread liquidity across every stock I like. If the market gives me ten decent setups, I don’t need ten positions. I just want ones where the narrative, trend, fundamentals and risk/reward are strongest.

The book’s recovery wasn’t one lucky trade, but a succession of actions after a succession of mistakes. I realized them, understood what needed to be done, and focused back on my system. I cut weakness, kept liquidity available, waited for the market to become “easy” again, bought retests and put size on the setups that deserved it.

The system

The last few months helped me reinforced my system.

Environment → strongest names → R:R setup → position size.

If the market confirms the thesis, I want to hold as long as possible, trim on extensions and add rather than search for a new position. Liquidity is always better spent on growing winners than new positions.

If the market doesn’t confirm, I cut weakness. I have no need to hold a stock buyers do not step up to buy where they should. There’s better elsewhere, or maybe it is time to slow down.

Nebius remains the prefect example to illustrate.

Breakout on clear volume marked “BO”. Clear extensions to trim marked “TP”. Clear W50 retests 40% lower marked “RT”. Repeat.

The system isn’t designed to eliminate losses. It is designed to make losses as small as possible and the wins as long as possible while increasing size on winners.

What the recovery actually proves

AAOI was a large contributor of the bounce, but the real contribution came from my own refocus into a winning system after having lost it in euphoria. Previous winners compounded while my refocus found new ones.

Nebius gave an excellent W50 retest. PATH and HLIT bounced really well. Akamai gave many buying opportunity to grow a position. The market started rewarding the setups that failed in July, and I became increasingly aggressive as it happened.

That is the part I care about.

The recovery wasn’t about becoming better at predicting what would happen. It was about recognizing that the environment had changed and adapting to it. When the market was difficult, I should have been smaller and more selective. When it became easier, I needed to do the opposite.

Press during easy times. Slow down during hard times.

It sounds obvious. It is much harder to actually do it when your portfolio is falling.

What happens now?

My stock picking book is back at ATH. But the return isn’t what matters most. What matters is that I lost the system during the drawdown, recognized it, went back to it and improved it.

I don’t need to predict the next correction. I need to recognize when the environment changes. When the market is difficult, I slow down. When it becomes easier, I need to press. Cut weakness, let winners run and concentrate liquidity where the market is proving me right.

I don’t need anything more complicated than that.

If you are interested in an actionable setup, this article details my latest position which is still buyable. If you are interested in the position planning, this article goes over the space sector bounce and two great opportunities. If you want to receive transactions live alerts and my buy target spreadsheet, you should consider the Founding Member subscription.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

Space is slowly becoming a new region to colonize, with almost everything still left to build. The next decade will be a race to establish infrastructure, secure access and exploit the region’s opportunities. Controlling space will become a source of military power, geopolitical leverage and, eventually, economic value. This is why the U.S. government is accelerating spending, missions and R&D programs to develop the technologies required for this new infra.

Space is becoming one of the most important strategic regions in the world.

We see it with RocketLab and accelerating demand for launches/services.

For Launch Services, demand is extreme. In Q2 and since the end of the quarter closed, we have signed more than $437M in bookings for Electron, HASTE, and Neutron. This includes a record $266M contract for up to 18 suborbital missions for the U.S. Space Force, our largest launch contract ever. Also, we have seen a massive surge in Space Systems contracts with more than $581 million signed in Q2 and post-quarter.

Q2 and the weeks after saw us awarded significant contracts across Space Systems, including a $397 million contract to build and launch multiple Flatellite spacecraft for the Space Force Space-Based Airborne Moving Target Indicator program. This program is a high priority for the Department of Defense.

With ASTS with governmental contracts for communication systems.

Meanwhile, the U.S. government customer has been a major focus for us, and we see great progress this quarter, both in terms of revenue capture and building the backlog. We drove revenue against several existing contracts and received three new contract awards.

With Firefly for defense and space colonization hardware.

After seeing our end-to-end capabilities, Carlos shared that our recent MoonFall program award is not only important for space exploration, it is also important for national security as future missions could collaborate with the Space Force.

Additionally, space is a national security priority, as shown by significant year-over-year increases in the U.S. Space Force budget, major conflicts around the world, and evolving missile threats. Market signals from customers point to orbital launch only getting more supply-constrained. Demand for government and commercial constellations and missions outstrips the rocket capacity available both in the U.S. and globally. Moon Base, national security, Golden Dome, and assured access to launch are driving significant industry tailwinds for space.

Or with Voyager for the same reasons.

that demand across defense modernization, national security, and the rapidly expanding space economy continues to accelerate

One company saying this is interesting. Many companies saying the same thing is information. It gives us our thesis confirmation: space economy is moving forward.

And after months of weakness, we’re finally seeing the other part of the thesis start to work: price action. Not everywhere though. The market is clearly making a distinction between space and defense. Most companies operate in both, but not to the same degree. For now, the market rewards companies with stronger exposure to defense.

That distinction is creating the setups I care about.

Numbers Confirm the Thesis

All companies I follow had strong quarters. Not just optimistic commentary, but actual revenue, contracts and backlog growth.

Rocket Lab delivered a double beat despite a negative revenue mix and payment timing, a record $2.36B backlog with 45.5% of it expected to be recognized over the next 12 months, and management guided to 10% sequential growth.

But it came with two negatives.

1. Neutron first flight was delayed, which means more R&D spending and a longer wait for commercial revenue. But management’s explanation makes sense: more work to reduce risk and scale faster after the first launch. Not thesis breaker, this is just classic space R&D.

   Production currently lines up with our target delivery of Neutron to the pad in Q4 2026. While the window for an end-of-year launch is narrowing, the work we’re doing now is about risk trading, balancing the timing of our first launch against how quickly and seamlessly we can scale our 10th launch.

2. Iridium acquisition is a question mark. Rocket Lab wants to keep Iridium’s core business while using its spectrum for new services, without giving details yet. I’d trust them - they have the track record on their side, but this is still an execution risk.

   We will expand upon it and scale it into untapped markets and pioneer new space-based services.

   It is through these growth areas that Rocket Lab will put Iridium spectrum to more effective use, extracting substantially greater capacity and throughput from the same finite spectrum allocation.

Rocket Lab is the only space pure-play I follow that has reclaimed its W50. I shared an alert with Founding Members ~$75.5 when the stock retested after earnings as I believe this was a clear overreaction. The position is up ~6%, and I’ll detail my view on it later in this article.

AST SpaceMobile also delivered a strong quarter. There wasn’t much that changed the thesis. Management continues to execute satellite manufacturing, launches and contracts with proper cash management, while preparing for future cash generation once the constellation becomes operational.

If anything, the bull case got stronger as they announced new contracts (notably with Rakuten Japan for a $1B market) from different customers and different use cases, including commercial telecom and government.

This follows continued work with FirstNet emergency and first responder networks in the U.S., with partner AT&T, and recent announcements with multiple governments through partners like Vodafone and Rakuten.

Redwire is perhaps the simplest example of what matters for these companies.

Space stocks move on contracts, execution and balance sheets. Successful operations and sufficient financing capacity lead to more contracts. More contracts lead to higher revenue and future cash generation. Which leads to a higher stock price.

Simple.

Redwire reported a reduced debt load, strong execution and record backlog.

Guess what? Stock went up. How surprising. More on this later as well.

Firefly delivered more of the same, expanding manufacturing capacities to meet demand, while winning more lunar missions, commercial payload contracts and Golden Dome-related contracts, without major delays.

It isn’t my favorite company or setup, but the numbers confirm the broader thesis.

Intuitive Machines is similar. The company is involved across NASA lunar programs, Golden Dome and other national security programs, while its backlog grew from $1.1B to $1.8B, and confirmed the broader thesis.

Combined with our existing leadership in lunar transportation and infrastructure, we believe these capabilities have expanded our addressable market from roughly $20 billion only a few years ago to well over $150 billion across civil, commercial, and national security space sectors.

Again, the important part isn’t one contract. It is the pattern. Demand is spreading across all parts of the space economy.

And then, we have Voyager. I’ve been watching this name for several months after it appeared on my screener earlier this year, even if I never spoke about it. Here’s a brief recap of their business.

Voyager is a defense technology and space infrastructure company operating across three complementary businesses: Defense & National Security, Space Solutions, and Starlab.

Defense & National Security supplies propulsion systems, communications, guidance, navigation, signals intelligence, missile-defense technologies and edge computing to customers including the U.S. Department of Defense, Missile Defense Agency, U.S. Space Force and major defense primes such as Lockheed Martin, RTX and Northrop Grumman.

Space Solutions provides the infrastructure needed to operate in space, including spacecraft communications, mission hardware, robotics, science payloads and mission services for NASA, international space agencies, research institutions and commercial space companies such as Blue Origin.

Starlab is Voyager’s long-term growth project: a commercial space station being developed as a successor to the ISS after its planned retirement ~2030. The project is backed by NASA and developed alongside partners including Airbus, Mitsubishi, MDA Space and Palantir, positioning Voyager to become a long-term provider of commercial orbital infrastructure.

Before this quarter, the problem was simple: Voyager was spending on R&D without evidences that it would translate into meaningful contracts and revenue. The market doesn’t like that, not enough to reward a stock.

This quarter changed the picture.

Voyager reported a record $336M backlog up 51% sequentially with $113M of new bookings during the quarter, and management raised guidance.

Perhaps the most significant takeaway is that demand continues to build faster than what we’re converting into revenue. As bookings consistently outpace revenue, backlog continues to expand, reinforcing our confidence that today’s growth is supported by durable customer demand rather than quarterly timing.

Looking at the balance of the year, we continue to expect revenue to accelerate through the second half, with approximately 40% of second half revenue generated in the third quarter and 60% in the fourth quarter.

Starlab also reached $500M in signed commercial reservations. Clear long term demand for a NASA funded project paid per milestones. One step at a time. This quarter was about de-risking the stock with market fit confirmations and new significant contract wins - notably a partnership for the Golden Dome.

The market loved it.

My Favorite Setups

Space companies have different problems from the companies I normally cover, and their stocks move on different factors. They are not simply scaling revenue and margins. They receive contracts, build hardware, launch or deliver it, finance the next stage and repeat the process.

Execution matters enormously. Delays mean more spending, more financing needs, less trust and potentially fewer contracts. And because most of these companies also rely on debt or equity financing, dilution is another risk for the stocks.

Their playbook is simple: execute, control costs, secure contracts and raise capital strategically when the stock gives them the opportunity; similar to neoclouds.

The market tends to price the sector together. When one company wins, investors often buy the whole group, because government spending is spread across suppliers to maximize the chances of success. That creates a lot of volatility from sector sentiment, contracts, delays, financing, dilution… which give very interesting buying opportunities.

This quarter’s earnings continue to confirm the sector thesis. More importantly, price action is starting to tell us which companies the market believes will capture the most value.

Right now, I see two clear setups and one additional name worth watching.

I’ll go through why, where I want to buy them, what would make me change my mind, and why I’m staying away from the rest of the sector for now.

Read more

Some names are already back near ATHs, some are making new highs, and every short-term support level seems to be bought while earnings are confirming that AI is only getting started, with the first meaningful data on inference and what enterprises actually want from compute.

We’ve already had great results, $AAOI up ~50%, $NBIS accumulate ~$155. $AKAM giving many buying opportunities ~$105 or $HLIT bought at $11.5 now up ~30% after excellent results.

This is the kind of environment where we have to press.

Some will say “you aren’t a genius, it’s just the market”. So what?

I’ll gladly take advantage of an easy market.

Every major AI compute provider has now reported. The message is clear: demand is massive, customers want faster and cheaper tokens, and providers need to maximize compute per unit of space and energy.

One company sits directly at the intersection of those, and fell ~20% after an excellent quarter because the market is looking at the results through one very specific lens.

This article explains why I think that reaction is wrong, what needs to happen for the thesis to play out, and exactly where, how and why I’m willing to take the risk.

Read more

Here’s what we’ll cover:

• Networking remains the strongest AI theme with Arista, Astera Labs & AAOI.

• SiPho demand continues to accelerate with Tower Semi, GlobalFoundries.

• The 800VDC thesis is facing execution issues with SolarEdge and Fluence.

• Cyber-security companies are the largest AI SaaS winners with Akamai.

• AI SaaS is proving its value with Palantir and Zeta.

I also discuss two investing concepts through Fluence and Palantir that I believe are must-reads if you want to become a better investor.

Before we begin, a quick reminder that I comment on the market almost daily in the Substack chat. If you’re only reading the articles, you’re missing a large part of the subscription where I share market views, price action, portfolio updates, alerts and much more in real time.

Now, let’s dive into this earnings season.

The Networking Bull Run

Arista Networks

Core position in Substack portfolio | Bought at $141 on 14/05/26 | +39.9%

A quarter ago, the market sold Arista on what I believed were overblown concerns.

My first takeaway is that the market is treating Arista unfairly. Yes, gross margins are a bit hurt with component prices and the current memory supply shrinking and yes that could be the case for years, but that will impact every hardware providers’ margins eventually and Arista’s management has always been first class, with premium operating margins which aren’t moving. Their wrong this quarter was to have a too honest management, but the truth is that they are a leader in the component with the most demand at the moment: XPOs.

I bought Arista today at $141. I only built a 60% position size for now, and will buy the full position if we go closer to its W50.

The Few Bets That Matter Substack

This quarter confirmed my view.

Today, management confirmed they worked hard on resolving those supply issues and increased their guidance to 40% YoY growth with stable gross margins for the year, stating many times that the market was focused on networking solutions for scale-in/across/out with their products being flagship products for today’s concerns - while they continue to develop tomorrow’s architecture.

I have never witnessed the combination of rapid innovation and scale deployment that we are seeing in AI networks.

To match our capacity with customer demand, we’ve increased both our manufacturing and distribution capacity.

Ken Duda & Todd Nightingale, Arista Q2-26 Earning Call

Networking demand remains massive because inference requires faster and more efficient infrastructure and Arista continues to execute better than anyone else in the sector.

Investing wise,

AsteraLabs

No position | Would buy on opportunities

The thesis continues to play out, even though I had to exit the position in January after it lost its W50. Back then, my thesis was simple:

The necessary next step is optimized, more efficient compute, to help providers deliver more with less. This grows revenues and margins by reducing costs & will reassure the markets about the financial soundness of [hyperscalers] expansion.

Maximizing compute per unit of space & energy is core for any provider’s profitability. Demand for the hardware capable of doing so will grow because providers have no choice but to buy it. They are in a race against their own leverage & need efficiency to generate cash as fast as possible.

The Few Bets That Matter

This quarter confirmed exactly that.

Revenues are accelerating, margins remain stable; the main driver is growing demand for scale-up AI infra. Scorpio X entered volume production ahead of expectations, while Scorpio P continues to expand.

On the AI fabric front, Scorpio delivered significant growth in Q2. High-radix Scorpio X-Series has entered volume production and will continue to scale materially in the second half of this year. We expect Scorpio to become our largest product family in Q3, which is one quarter ahead of our prior expectations.

We are now shipping multiple configurations of Scorpio X-Series into scale-up applications to our initial customers, we remain on track to ship Scorpio X-Series to additional customers by year-end. Scorpio P-Series also continued to expand across multiple customers, with new shipments to hyperscalers and AI infrastructure providers. These programs are currently in pre-production and are expected to ramp more meaningfully in 2027. We are also closely engaged with our customers to support their future system architectures with purpose-built AI fabrics, including next-generation PCI Express and UALink protocols. Upcoming AI systems will employ more complex switching topologies to enable larger XPU cluster sizes and higher bandwidth. These designs are expected to drive a material increase in our silicon dollar content per XPU.

Jitendra Mohan, Q2-26 Earning Call

The thesis was about AI infrastructure becoming increasingly optimized. Nearly a year later, that’s exactly what we’re seeing - and it plays right into Astera’s hand.

Management also confirmed the optical roadmap remains intact, with NPO expected around 2027 and CPO further out, with products already being developed for both transitions.

Investing-wise,

Applied Optoelectronics

Core position ~20% of book | Bought at $91 on 03/08/26 | +42%

I am not one bit surprised by this monster quarter. It’s the continuation of what we already knew, and what the rest of the optical supply chain confirmed this earnings season. Hyperscalers are upgrading bandwidth, which means upgrading connectivity. That’s exactly where AAOI benefits.

The take away of this quarter is simple: demand outpaces supply, like for many other hardware sector.

Demand to support next generation AI infrastructure remains so robust that our near-term revenue is bounded almost entirely by production capacity and key component availability.

During the quarter, we saw continued robust customer engagement around our 800G and 1.6Tb products. We saw a strong volume ramp of our 800G product in Q2, which more than doubled sequentially. Forecast demand continue to outpace our production capacity through mid-2027.

Looking ahead, forecast demand for 800G and 1.6Tb modules are projected to continue to exceed our production capacity through mid-2027.

Thompson Lin, Q2-26 Earning Call

The message is clear: demand is not the question, capacity is. AAOI is expanding its fabs to meet it, partly financed by customer prepayments as hyperscalers race to secure supply.

Their capacity is expected to increase from ~200k monthly units today to 650k by year-end and ~900k next year while product mix should continue shifting toward higher-value modules. More units and a better mix means into higher revenue, margin and cash generation.

Everything is about execution from here. If management continues expanding capacity while meeting demand, the market will continue to reward the stock.

Investing-wise,

Photonics Foundries

Before diving into GlobalFoundries and Tower Semiconductor, let’s take a step back.

Both companies are expanding production to meet accelerating demand for silicon photonics, primarily driven by NPO deployments in AI data centers. That highlights one of the most interesting AI supply chains to own today:

Soitec (SOI wafers) → GFS/TSEM (SiPho manufacturing) → Arista (NPO and networking).

Marvell also plays an important role through its engineering.

Looking at the supply chain instead of individual companies reinforces my conviction in Soitec. If the two largest silicon photonics foundries are expanding capacity, demand for SOI wafers should naturally follow.

Global Foundries

Sold at $68.8 for a 63.8% gain | Could buy on opportunities

The thesis continues to play out. Demand for silicon photonics is accelerating - exactly what we discussed with Soitec a few days ago, led by pluggables.

Within this end- market, we saw strong customer demand for our silicon photonics and silicon germanium offerings. In both of these high-margin technologies, we’re ramping capacity and making the necessary investments to unlock increases in demand indicated by our customers.

Given the accelerating demand outlook from our customers, we now expect to achieve full-year 2026 revenue growth in the range of 50%-60% for our communications infrastructure and data center end-market, up from our prior expectations of high 30s percentage year-over-year growth, which we believe is an early indication of the long-term growth opportunities ahead for GF in this end- market.

Sam Franklin, Q2-26 Earning Call

Demand is no longer the question. GlobalFoundries is expanding capacity because customers are asking for more, not because management is hoping demand will arrive.

The challenge is that AI represents only ~16% of total revenues. Communications infrastructure and data centers are accelerating rapidly, but automotive and mobile remain much larger segments and continue to weigh on overall growth.

Management confirmed they’ll continue to expand capacity for SiPho, boosted by a $300M award from the U.S. Department of Commerce, and are planning to increase prices.

With respect to pricing, we’re encouraged by the improving industry dynamics as well as the evolving mix of our business towards highly accretive technologies. In addition to these positive mix shifts, in the second quarter, we implemented pricing increases in partnership with our customers across several technology corridors. Following the satisfactory conclusion of these customer conversations, we expect the pricing adjustments to be reflected in revenue commencing in 2027. The magnitude of these pricing increases varies by end market and technology and contemplates the differentiated value we provide, the ongoing supply and demand dynamics, and the inflationary absorption across our industry in recent years.

Sam Franklin, Q2-26 Earning Call

The pricing discussion may be even more important. Growing demand is translating into pricing power, which should support both revenue growth and margins over the medium term.

Quantum also received attention with the new $375M U.S. grant. It’s a positive but not something I’d include in an investment thesis. Great upside if it works, impossible to model today.

Investing-wise,

This portion is reserved to paid subscribers and goes through how to maximize returns. Upgrade to get full access. Free content resumes below.

Tower Semiconductor

No position | Could buy on opportunities

TSEM is showing the same dynamics: accelerating demand and capacity expansion, without the drag from legacy businesses. As a pure AI foundry, its results reflect the strength of demand for both compute and power semi.

The company reached a $2B run-rate, secured prepayments for 2027 capacity driven by NPO demand, guided for faster acceleration into 2028 and is expanding all of its fabs to meet that demand.

Silicon photonics revenue itself increased by over 60% quarter-over-quarter and over 270% year-over-year, hitting a Q2 annualized run rate of over $680 million, targeting a $1 billion run rate in Q4 2026 as the previously announced capacity expansion continues to be qualified.

During this quarter, we experienced increased demand from existing customers and also saw very strong new customer acquisitions across our power portfolio. This momentum is primarily driven by growth sectors where power density and thermal efficiency are mission critical.

Russell Ellwanger, Q2-26 Earning Call

Hard not to be bullish.

Investing-wise,

Batteries On Fire

Fluence

Sold at $17 for a -12% loss | No plans on buying

We’ll use Fluence as a practical lesson on one of the most important investing principles: respect the market.

Fundamentally, everything looked fine. Backlog continued to grow, with a record $1.44B intake, including the $850M order from the first MSA signed a few weeks ago to supply one hyperscaler with behind-the-meter battery systems.

That order confirmed my original bull thesis. Yet the stock still fell 30% after earnings.

We should have seen it coming, even without knowing why. The stock had been warning us for weeks. It lost its W50, broke every meaningful support and never reclaimed them. That kind of persistent weakness usually means one thing: someone knows something we don’t.

Millions of participants interact in the market every day, and liquidity itself carries information. We didn’t know why the stock was weak. We only needed to know that the people moving the most money weren’t buying - they were selling.

Eventually, we found out why.

The company missed revenue expectations and lowered guidance due to execution issues, with production delays at two new manufacturing facilities.

Q3’26 revenue $90 million below expectations due to production delays at two new contract manufacturing facilities:

• Three-month delay in ramp of Houston enclosure facility, now initial production

• Initial quality issues from international facility, projects now reworked and being shipped

Fluence Q2-26 Slide Deck

The thesis was never simply that hyperscalers needed energy. It was that they needed it today. Bloom Energy is being rewarded because it overdelivers, today. Fluence did not and the market immediately repriced that execution risk.

The market is often excessive. It rewards companies that consistently overdeliver and punishes those that disappoint because it assumes those trends will continue.

I’ve been criticized for selling while “the thesis hadn’t changed”. I disagree. The thesis had changed, we just didn’t know it yet. The market did.

That’s why I respect price action.

SolarEdge

Sold at $42 for a 23.5% gain | No plans on buying

This is a different story from Fluence, I believe the market is overreacting.

We already knew H2-26 would be weaker because management guided for it last quarter. The surprise wasn’t the direction, it was that the core business deteriorated more than expected.

Meanwhile, development of the SST platform for 800VDC AI data centers continues as planned, but commercialization is still years away.

We remain focused on our next planned milestones, getting to a working system in our lab by the end of this year, followed by pilot installations in 2027 and volume shipments in 2028.

Just like Fluence, price action warned us before earnings. The stock lost its W50, failed to reclaim its previous breakout level and was rejected every time it tried to recover its daily averages.

I’ll keep following the company closely, but I won’t do anything until the market starts confirming the story again.

The Biggest AI Winner

Akamai

Core position in Substack portfolio | Bought at $110 on 28/07/26 | +22%

Momentum continues to build.

Cloud computing and cyber keep gaining traction. It isn’t translating into accelerating growth just yet, but management slightly raised guidanceand shared positive comments fo FY27 as demand continues to strengthen.

A few weeks ago, the market sold the stock on higher CapEx and it bought the exact same story today. Clear shift in perception as CapEx is tied to accelerating demand, just like for Google or Micosoft which becomes extremely attractive for investors.

Customers that added or significantly expanded their WAF usage in Q2 included one of the world's leading commerce sites, one of the world's leading automakers, and one of the world's largest banks in a $14 million upgrade to have Akamai secure all of their applications.

As a result, all of our GPU capacity is completely sold out. Therefore, driven by the $600 million new customer win we announced today and a very robust pipeline, we expect to invest up to $500 million in CapEx to replenish and expand our GPU capacity.

Tom Leighton, Q2-26 Earning Call

AI is a clear tailwind for both Akamai’s cloud and security businesses, and we’re finally starting to see it with customer wins and higher guidance.

Investing-wise,

The SaaS Renewal

Palantir

Sold ~$170 for a ~500% gain in late 2025 | Could buy on opportunities

We could spend pages on the numbers, but there isn’t much to say. Palantir keeps doing what it has done for years: accelerating revenue, expanding margins and proving that AI can be transformed into a valuable product. Fundamentals are simply outstanding.

I sold ~$170 because the stock was extended and valuation left no room for error - although there was no error yet.

Ironically, Palantir is one of the best examples of what value investing actually is. Many investors define value as buying companies trading at low multiples.

The market doesn’t; it prices future cash generation.

When a company continues to accelerate revenue, expand margins and prove product-market fit, high multiples are justified because the market is pricing future cash flows and success, not today’s earnings and failure. Like Fluence versus Bloom, the market prices continuity.

Back to earnings, the most valuable takeaway from this quarter wasn’t the numbers but the confirmation of where enterprise AI is heading. Commercial demand continues to explode, with AIP revenue growing 149% YoY and full-year guidance raised to 135%.

Why? Because enterprises don’t simply want AI. They want AI that solves problems, and that’s what Palantir delivers. Many companies are still experimenting with direct LLM usage, paying for tokens for no meaningful value. Palantir takes those models and turns them into operational software that improves workflows, decision-making and productivity.

That’s why I believe AI SaaS will become one of the biggest winners of this cycle. Just as they optimized business software over the last two decades, AI SaaS will optimize how enterprises actually use foundation models.

Palantir understood that before almost anyone else.

In contrast, enterprises that are not using Palantir are seeing their token meters spinning endlessly just to get slop without any correlation to value. This token model may be working for the labs, but it is not working for anyone else.

On our side of the chasm, what enterprises demand is AI sovereignty, owning the operational definition of the data, logic, actions, and security of their enterprise.

We already learned that different AI workloads require different hardware, specially optimized for them. I see no reason why software will be any different.

Foundation models provide intelligence. AI SaaS provides optimization. That’s where enterprise value is created.

The assumption that the frontier is actually the best performing is just not borne out in practice. Within 24 hours of bringing Nemotron Ultra into our stacks, we found five production tasks where a standard Nemotron Ultra model without post-training beat frontier models. This underscores that a handful of common benchmarks can be gamed.

The reality is that the market has created far more intelligence than it has converted into value.

Investing wise,

Zeta

Sold at $22.25 for a 17.1% gain | Could buy on opportunities

Zeta confirmed my original thesis that AI SaaS should produce clear winners in the short term: companies selling proprietary data and those using AI to optimize business workflows. Broader AI services for consumers will take longer to mature.

Zeta sits firmly in the first category, delivered another double beat, raised guidance and confirmed that demand for its data platform continues to accelerate.

The initial sell-off after earnings was due to incorrect numbers published by financial media, triggering algos selling for no reasons. The stock recovered, as it should have as management highlighted growing adoption of the platform and confirmed that its partnership with Palantir already resulted in “multiple agreements for combined sales”.

That reinforces my broader AI SaaS thesis.

Investing-wise,

In conclusion,

This earnings season reinforced one message above all else: follow the winners. AI hardware continues to lead both fundamentally and technically.

Arista confirmed demand for pluggables remains strong while preparing for NPOs. Tower Semiconductor and GlobalFoundries confirmed accelerating SiPho demand and capacity expansion. Astera reinforced that optimization remains one of the biggest bottlenecks for AI infrastructure. Nothing points to a slowdown before at least next year.

We’re also starting to see a second wave emerge: AI SaaS. Palantir continues to prove that enterprises don’t just want AI - and fail using LLMs; they want AI that creates measurable value. Zeta confirmed the same trend through accelerating demand for high-quality data access.

On the other hand, energy reminded us that a good thesis isn’t enough. Execution matters. Both Fluence and SolarEdge showed that the market has very little patience for delays, even when the long-term opportunity remains intact.

That’s why I continue to focus on strength.

Follow the sectors where fundamentals, execution and price action all point in the same direction.

Today, that’s AI hardware first.
AI SaaS is starting to join.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

We finally have some answers on what happened, I’ll go over them today. The positive is this drawdown allowed us to identify strength and weakness in the market, and we now can focus only on the former for the rest of the year.

I’ve learned a lot over the last weeks; some things were well done, others weren’t… Looking at my portfolio, there was more wrong than good, and while some could argue that this sell-off wasn’t organic - as in it makes it acceptable to have such a drawdown, I won’t; the “why” doesn’t matter. Manipulation is always present, whether you see it or not. Complaining about something you can’t change is a waste of energy, and using it as an excuse is hypocritical.

What isn’t a waste of energy is improving, so that you can perform during this kind of inorganic sell-offs.

Portfolio, Gift & Lessons

My portfolio is now up “only” 21.7% YTD. I should be happy; being up after such a drawdown while holding growth names with leverage is something. But going from +195% to +21.7% in two months is unacceptable.

I knew the breather was coming and still had such poor risk and margin management that I almost gave up all my performance. I also shared content I’m not satisfied with. And I won't hold on to a paid subscription while not being satisfied with my content and performance.

For these reasons, I extend all current paid sub by two months. If I don’t perform according to the standard I set for myself, you guys have no reason to pay for it.

Some things were well done and I hope useful. I cut lots of weak names, which ended up falling much lower - despite lots of criticism, and still trade largely below my selling price. I found great opportunities at great prices, notably with Amazon, and shared great entry prices for Founding Members’ on Nebius (~$155) while holding on names like Arista, Soitec, and Silicom which were clearly strong.

Everything wasn’t a disaster. My portfolio is still green, the stock-picking portfolio is up 46.42% YTD; strong, considering most of the social media sphere, but having my own performance below it means I did something really wrong.

The year isn’t over, but I hold myself to a higher standard.

As for the learnings, I already shared a few of them. This kind of drawdown won’t happen every year as manipulation to this extent requires specific conditions. But harsher rules on how and when to scale margin are now in place, with much stricter risk control. Margin scales only on clear uptrends, not before.

Overall, things could have been worse. They have been for many. I haven’t done perfectly, not even well, but enough to stay alive and remain green YTD. I now have a great watchlist of strong assets, an improved system, and an even stronger resolve to make a killing.

These weren’t easy weeks. But chances are we’ve seen the worst.

Leopold’s Degenerate Gambling

Talking about staying alive, here’s one who got a margin call and lost all of his clients’ money - despite what you can read around.

I will only go briefly over the story and market’s mechanism; what happened briefly, and why it impacted us so harshly. If you want a punchier and more detailed version - also more complex, Irrational’s your guy.

Leopold - the social-media-praised and famous AI investor, was on 4×1 margin and got a bank call on Wednesday as his assets no longer covered his margin requirement. Before that happened, he tried its best to cover himself and sold his assets to reduce margin, crushing prices until the inevitable happened. Citadel ended up buying what was left on his book at a great discount when he had no other solution but to get out; this is why the market bounced on Thursday and Friday.

That’s the overall story. But the speculation goes further.

Leopold’s problem was that he let in some people he probably shouldn’t have. Those had regular and unrestricted views of his assets and margin levels and saw the risks he was taking.

A logical assumption would be that most of the drawdown we saw was an organized attempt to trigger Leopold’s margin call: shorting, selling, and no buying. This dip was too violent to be organic, with no bounces at all or extremely small ones as Leopold kept selling while others kept putting pressure on his positions. When he finally blew up, Citadel got a great discount, closed its shorts alongside everyone else, the market ripped again. Some of that sell-off was organic, but the violence of it probably came from this mechanism.

Wall Street isn’t for the weak.

This could be the scenario of a movie - might be one day: the fallen prodigy. Even if this is only speculation, make no mistake, this is how Wall Street works. If one can make billions and destroy two or three others on the way, he’ll do with a large smile on his face. People at Citadel must be drinking some pretty expensive champagne this weekend.

I’ll send my script to Netflix; what matters is that this explains the violence of the late drawdown and why no bounce held at all despite positive news (col CPI/PPI, no rate hikes, great earnings…). The only thing that matters in the market is liquidity. Narrative second. Fundamentals only third.

And liquidity was focused on destroying Leopold.

Bottom or Bubble?

Talking about market mechanisms, chances are we’ve seen the bottom. This doesn’t mean we’re back in an uptrend, but means we’re good to buy again. And I believe the next few days will be key to the rest of our year’s performance.

Most of the selling pressure is behind us. Leopold has nothing left to sell, Citadel is now holding the chips and might - should, sell some, but not everything considering Leopold’s portfolio quality and latest earnings. The portfolio might be a bit outdated, as he had to sell some holdings, but its quality is undeniable.

At this point, I see two major liquidity pressure points left:

• Citadel and traders offloading some positions on pumps, which could put some pressure on these names - and the global growth market.

• Retails and funds now buying back great fundamentals on great earnings and a healthy environment - except for Iran, cleared of most leverage and risk now that valuations have taken a large dump.

I’d expect the second to be the strongest, but I’d also expect the first to give buyers a a hard time short term.

We should be moving from a downtrend to a fight between sellers and buyers - a range. I struggle to see sellers being stronger now considering the environment, but we shouldn’t rule it out. And I struggle to see buyers overwhelming sellers right away, but that is more likely than the contrary.

My expectation remains a range, as it has for weeks. But I am now more confident about the outcome direction of that range - higher, and its length - few weeks.

We are going to see a bubble, but this isn’t it. A bubble is built on endless optimism, and blowing up a leveraged fund by selling amazing earnings showing clear ROI from AI, revenue acceleration, and margin expansion isn’t really what endless optimism looks like.

Factually, yes, the AI build-out will slow down, models will evolve, and DeepSeek- or Kimi-like improvements to models and inference will reduce compute needs, and we will find a balance. Nothing points to it being found yet; everything points to the contrary, and the market is clearly not overly optimistic here.

Not every company is cheap, even today. But most fundamentals are improving so fast that it’s impossible to know what tomorrow will bring, especially considering the technological transformation we’re going through. And as the market isn’t pricing endless optimism here, I don’t understand why anyone would call this a bubble. I call it a perfect environment to focus on strength and hope for this bubble to form, for real.

The rest of this write-up will go over my current portfolio, stock picks, and how I’d play each name now that we have more clarity. I will also go over my curated watchlist now that we can clearly identify strength and weakness, and again detail how I’d play each name from here. Some extremely attractive risk reward are showing up on the market on very healthy companies.

This is the time to refocus, forget about the last few weeks, and get ready for the next.

Read more

If you listen to calls, this is what transpires; not only that demand for compute remains supply-constrained, but that AI is delivering ROI. Google showed this, in terms of raw compute/intelligence renting but also in terms of usage, with coding or advertising as the most impacted use cases for now, with many more to come.

The conclusion is straightforward for a business owner.

Spending generated CapEx and new AI usage emerge as capacity and pricing evolve. So they want more, to make more. Math is easy; you buy something at $10,000, have it refunded in ~3 years for a lifetime of ~6; half of its lifetime is pure profit. Why stop?

My original thesis shared ~a year ago was about optimization as a go-to for compute and energy. This is what continues to transpire today with continuous need for both while the market prices skepticism.

This week’s earnings review should confirm that we don’t need that much skepticism, and go over leaders’ price action with actionable insights on how to play a potential bounce coming in the next months.

AI Has ROI

Let’s start with Meta,

I’ve been very bullish and continue to be, although I had to close my position as the stock lost its support and fell further post earnings, despite a great print, confirming my bull case and my assumption on why the market doesn’t like this stock.

Meta confirmed AI is the source of their financial and business improvements, but had no specific “AI ROI and revenues” line on its statement with no announcement on compute rental, which continues to make that spending look inefficient - even though it isn’t.

The opportunity in front of us is massive. First, we are now at a point where our investments in AI are accelerating every major part of our core business. They're improving the experience for people using our apps, driving better performance for advertisers, and helping our teams build new experiences and ship faster.

Increase in both advertising efficiency and pricing is due to AI, combined with increased spent time apps.

This is where compute goes, and it shows in Meta's revenues and margins which continue to be healthy considering the company's size, with margins falling due to two one-time expenses, which made the quarter look "bad" with a 6% impact on operating margins.

Q2 total expenses were $42 billion, up 55% compared to last year, and included $2.4 billion in charges related to legal proceedings and $1.2 billion in severance expenses in connection with the May 2026 headcount reduction.

We’d still have a decrease due to higher depreciation/costs, which the market would still dislike, but these levels remain really healthy.

The real battle for Meta remains perception; they need to show that their CapEx does more than “just improve” their core business. They need to show ROI, and confirmed during the call that they will rent compute/intelligence directly to third parties, but did not share any details except to say that we’d hear more about it soon.

This is the catalyst - or so I believe. We’ll talk about Microsoft and Amazon, but when seeing these companies’ results, after Google’s, it is clear that renting AI compute generates ROI and the market is starting to reward it. I see no reason why Meta doing the same wouldn’t trigger stock demand.

We've built our API. We're rolling out business agents. We're getting a lot of offers for compute at a significant premium over what we paid for it, and we have more coding and productivity tools on our roadmap as well.

This is also why management contracted a $14B joint venture with BlackRock for a 1GW AI data center in Texas, owned at 80%/20% by BlackRock and operated by Meta. More plans for more compute - with some clear financial engineering as this will stay out of Meta’s balance sheet due to the structure.

So we are starting to see a few things:

1. Hyperscalers are using more financial engineering; leverage is starting to increase for the entire economy.

2. Banks and external investors are starting to put skin in the game, with BlackRock here committing cash and raising debt, after Google’s $10B raise from Berkshire.

3. No slowdown in buildouts, and institutions are ready to push harder; still financed “healthily” for now with reasonable debt, FCF spending, or dilution, but we start to see an industry raising risks.

Risk which delivers growth and ROI is good risk - Amazon confirms this; issues start when the balance shifts.

The question left isn’t “does compute have ROI?” but “does AI have ROI?”. This won’t be answered by hyperscalers, but by SaaS, as they are the ones consuming tokens to deliver services. Considering the amount of tokens they’re using, I’d imagine they see ROI from it but we are yet to see the real numbers.

Investing wise,

The Compute Kings

Google delivered impressive results last week; this week is about Amazon and Microsoft. You should start by Andy’s laïus on CapEx and ROI, as he kinda treated Wall Street’s analysts like kids and explained them how investing for a company’s future worked.

And the numbers prove him right, as the company continues to accelerate, driven by its cloud/AI/AI cloud business.

Note that Amazon received two one-time inflows - one regarding tariffs refunds for $1.8B combined, so operating margins should be ~12.8%, still extremely healthy.

I’ll start with AWS, which is booming right now, and I’ll share the numbers, what we think is going on, and why we’re enthusiastic about the ROIC equation, even with heavy CapEx the next few years.

First, the numbers. Revenue growth of 36.7% year-over-year, accelerating for the fifth straight quarter, our fastest growth in 18 quarters back when AWS was less than half its current revenue size. We added over $4.6 billion in revenue quarter-over-quarter, about 80% more than our largest increase ever. Our backlog stands at $496 billion, growing triple digits year-over-year.

AWS is now a $169B annualized revenue run rate business, which, for perspective, would place it 24th on the Fortune 500 list if it was a standalone company. Our chips business now has an annual revenue run rate of over $25 billion, growing triple-digit percentages year-over-year. Our AI revenue run rate climbed significantly QoQ, and is now also over $25 billion, growing triple-digit percentages YoY.

Revenue growth continues to accelerate, margins to expand, and advertising - due to AI, is also accelerating. Clear AI ROI from rental to usage, and Andy explains why the cycle is far from over, despite Wall Street skepticism.

In addition to leading model building and inference services, customers need easier ways to build, run, and leverage agents. For example, even after you've built an agent, you have a lot of muck to worry about. A production agent needs somewhere secure to run, memory so it holds context, an identity so it can act on a user's behalf, tools and data to connect to, and a way to watch what it's doing once real traffic hits. Stitching all that together reliably is hard, and it's stalled many production deployments.

Amazon will increase its CapEx once more this year and probably for the years to come, stating that demand for 2028 is “striking”. Will the market hold, and will this convert into real revenues by then? I do not know, but I don’t really need to. Those are bullish signs and the market should continue to price optimism, not because we are in a bubble and investors are stupidly bullish, but because data is bullish with no end in sight for concrete ROI. That’s all we need to know short term.

Regarding current position,

Microsoft is more of the same. I could go over each number with QoQ comparisons, but the conclusion would be the same: more demand, increasing margins, clear cash generation a few years after CapEx cycle with longer lifespan for their hardware.

Once again, the only question left is: does AI have ROI?

I still don’t have an answer to that and still don’t need one yet. What I need to do is gauge the market’s optimism, have a tight risk management, and be involved on the verticals it likes, while leaving behind the rest.

I had a position on Microsoft but it hit my stop loss. Thesis was on point, but timing was wrong; it happens sometimes. The demand zone was clear, and I marked it a few times already.

Now, one position stopped doesn’t mean we shouldn’t try again when conditions get better, and they now are. The stock hasn’t pushed above its current high; these last two days were a great sign, but I’d expect the market to give the best opportunities in the weeks to come, not just yet.

In MSFT’s case, this would be a retest of daily averages, with anything below $450 being great value, depending on how aggressive you want to be, with the W50 sitting at $430.

This stock hasn’t been the strongest these last months, but as the market is rewarding compute lately and with clear ROI from CapEx, Microsoft - almost a cloud pure-play, could become a leader in the next months.

In brief,

The takeaway is extremely simple. Compute demand is through the roof, supply limited for some time according to every major compute provider, delivering ROI, and they are all looking to build more. Not really bearish…

Now that we know that, the logical sector to benefit the most from this is AI hardware.

Silicom Is Cheap

I am not one to write overly confident bullish articles, nor to be cocky. But it becomes hard not to be when it comes to Silicom.

Double-beat with expanding operating margins, raised FY26 guidance from ~$82.5M to $94M - a 14% increase from last quarter, guiding Q3-26 to 66% YoY growth, and non-GAAP profitability H2-26 - earlier than I or management expected.

This trajectory was our exact bull case, coming from increasing demand for optimized networking solutions or inference, but also classic networking which requires to be upgraded in an AI world.

Management expected to win between 7 & 9 designs this year and won 7 in six months. The pace of demand from new and current clients, for new and current designs is accelerating, mainly for AI.

Why? Because there are no alternatives to FPGAs to improve networking connectivity. Because adaptability is key in today’s hardware world as everything changes so fast. Because constructor lock-in is a drag in many situations which require optimization.

if you're doing an ASIC, you are locked down for many years, and if you're doing it on FPGA, you actually can update all the time, and as models progress over time, you can actually take all the new goodies that you have and all the industry that is doing smart things every day and put it into your FPGA and actually run models quicker and better than what you did yesterday.

Seeking to move away from vendor lock-in, the customer decided to replace its existing proprietary switches from an incumbent industry leader with Silicom's open white label switching solutions.

each successful win opens the door to the next, with satisfied customers coming back to us for additional products and additional use cases.

This is Silicom’s timing of opportunity, exactly as expected.

On the AI inference deal, We knew Silicom was in discussion for two products (an AI-NIC solution and an FPGA-based solution already in PoC). The question on my last write-up was about which of those converted to the design win and production orders announced last month.

Additionally, we successfully customized an AI-NIC solution to meet the customer's specific needs, delivered the first unit to the customer evaluation, and are preparing for initial deliveries of this customized product per purchase order received from the customer, a leading AI-Inference ASIC and infrastructure vendor.

It seems the production order comes from the AI-NIC solution, not the ongoing PoC for the FPGA product. This order is expected to generate between $3M & $4M for 2026, with 2027 “much higher” while the “multi-thousand units” potential FPGA order hasn’t yet been confirmed. It might never be, as the PoC could be unsuccessful, but would be net additive if it were, while management could be in discussion for at least one more (new?) AI inference product.

In parallel, we are expanding our AI-Inference product portfolio, and based on orders secured, we are now developing a completely new bespoke inference-specific solution.

This comment could also be about the FPGA solution, so I’ll keep my expectations low and assume the AI-NIC solution sold to an ASIC manufacturer was converted into a production order, while the FPGA-based solution is still ongoing PoC. This situation is already extremely positive as selling NICs to an ASIC manufacturer largely confirms my bull case and the importance of NICs for inference already.

I also want to note Silicom implication in Quantum cryptography, which will require specific hardware for which they’re already in discussion with clients. This is another vertical that will drive growth in the future, although AI remains the #1 opportunity.

Increased revenues, increased demand, involved in AI inference, improved margins, profitability expected in the next quarters. The market prices stock based on potential future cash generation, which comes from accelerating growth, expanding margins, or both. Silicom is combining both, just like Lattice - which I already took as an example. The market rewarded its stock with an optimistic 30x sales.

Today, I believe Silicom’s only bear case at this price - besides black swans, is visibility. For a stock to rise, others need to buy it. For others to buy it, they need to know it. How does that happen?

• Social media awareness.

• Analysts’ coverage.

So far, only a handful on X talk about this name and one analyst covers it with a $60 price target shared before earnings - probably going to raise it after such a print.

As for investing,

Axt Inc

Great earnings as well and a 64% performance in two days… Just like Soitec, AXT is one of the best photonics play as they make the wafers for lasers; a key player very high in the supply chain, which confirmed the same kind of trend:

Continuous growth acceleration and margin expansion with massive demand, above current supply, which justifies expansions, while less than a handful of companies capable of selling comparable wafers.

The company signed a contract with Coherent, another fundamental confirmation, and shared that Chinese demand was accelerating which is great as it comes with less geopolitical tension - AXT and its producers are located in China and need permits to export outside of China.

One of the best photonics plays on the market, with an extremely interesting price action.

Bloom Energy

Stock price follows future potential cash generation, which can happen two ways: increased revenues or expanded margins. Just like SIlicom, Bloom does both at the same time and confirmed the trend was strong and for longer.

Energy was and remains the main AI bottleneck, and both storage and behind-the-meter generators are the new go-to solutions for the years ahead.

New customers are arriving at a faster pace than ever. Landing a mega customer used to take us years, a moat that protected the incumbents. Now, the proven success of our technology, coupled with an increasingly urgent need for efficient, clean, and reliable power, has collapsed the time from first engagement to first order. Because we can book, ship, and convert orders to revenue inside the same fiscal year, that demand shows up in results now, while also adding to and diversifying our backlog. Notably, just this year, several customers who had alternative solutions in place abandoned them and came to Bloom.

I see Bloom exactly like I see AXT: extremely interesting company and price action.

There are no questions of what Bloom is, and can become.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

I will detail Google’s earnings and what they mean for others, Soitec’s momentum and catalysts, ServiceNow as a proxy for SaaS and AI demand, and some comments on TXN and STMicroelectronics for power and analog semis.

The AI King

Incredible earnings, this quarter was a confirmation of all layers of the AI bull case. The market’s bear case for quarters was always ROI. Google answered.

Continuous acceleration for a company with $100B+ quarterly revenues, margin expansion - stabilization QoQ at healthy levels, and increase in operating cash flow, net revenues and EBITDA.

Whichever metric you use to measure AI’s impact on the business, you have only one possible answer: a positive one.

Be mindful of net income as it includes their P&L, rocketing after SpaceX IPO; that’s not cash generation, just mark‑to‑market P&L they include in their statement – and EPS which also came with a massive beat kind of “forged” by that.

Still, a remarkable quarter with AI impacting all of their verticals.

Most segments’ revenues are stabilizing, but stabilizing multi‑billion‑dollar revenue streams at double‑digit growth for years isn’t a small feat, and AI is the reason for this stable growth.

Next, the Gemini app, which now has 950 million monthly active users with daily active users tripling in the last year.

Those who adopt our AI-powered campaigns like AI Max or Performance Max, see an average of 15% more conversions or value on Search at a similar ROAS. AAA Auto Club Enterprises used AI Max to personalize creative assets and capture growth from increasingly detailed insurance searches. This led to a 17% improvement in conversion volume and an 11% decrease in cost per lead.

To attract high-value shoppers across North America and Europe, outdoor brand Arc'teryx used Demand Gen to achieve a 70% better return on their ad spend compared to other paid channels.

Kate Spade reached Gen Z through a first-of-its-kind YouTube creator campaign and partnered with creators Ellie Thumann and Hannah Meloche. By leveraging multiple format videos on YouTube, the brand drove a 3.25% brand lift in purchase intent.

I won’t go over all the quotes; you get the gist. AI also powers internal work:

As just one example, a team in Chrome is now on track to accelerate delivery by 8x , compressing a two-year timeline into three months through model-driven refactoring.

This confirms most bull cases as it shows real AI demand and valuable use‑cases for companies around the world.

The beast is Google Cloud, which closed with a record ~$514B backlog for which 50% are to be realized in the next two years, with strong growth and operating margins expanding from 20.7% to 35.6% YoY.

Rapid - and unperfect, math would suggest a ~65% YoY growth for the vertical NTM with ~$45B of operating income with today’s margin. Backlog isn’t perfectly regular so that' won’t be exactly the case, but it gives you an idea of the potential and what GCP has become, again thanks to AI. These aren’t “ifs”, but real numbers behind the AI infrastructure story and CapEx.

The market was doubtful on ROI; that should answer…

Management also confirmed what we knew: they are supply‑constrained, said they’ll use third parties to meet demand waiting for their own infra. This is extremely bullish for Microsoft, Amazon, Nebius, CoreWeave, Oracle, xAI & co as it means one of them will have a juicy contract, but also that demand for their product continues to increase.

We’ll see rental contracts in the coming months.

Demand for our models is translating to strong token usage across developers and enterprise customers, and we continue to be supply constrained, a sign of momentum and rapid adoption.

Sundar Pichai, Google CEO, Q2-26 Earning Call

Now, where market gets spooked, despite the numbers and obvious trend.

Google confirmed AI has a meaningful ROI across all its verticals: improved internal workflows, improved advertising revenues, time spent on apps, massive demand for direct compute and TPUs. What AI touches, AI enhances and it comes with higher revenues, margins and cash generation.

Google wants more compute, which means more spending. This quarter was Google’s first‑ever negative FCF; they increased their CapEx guidance to ~$200B and expect even more in 2027.

Finally, we expect the free cash flow will remain under pressure, driven by our investments in technical infrastructure, which enables us to capitalize on the AI opportunity and continue to drive attractive returns.

…we continue to expect our CapEx to increase significantly in 2027 and will provide more details at a later date.

The market was worried about ROI. It had an answer. Now it is worried about CapEx funding methods as it comes with negative FCF and dilution, even if spent on AI, a technology with back tested ROI.

I’m not sure what this means for Google’s stock in the short term; what I know is that it isn’t cheap and is extended, so I wouldn’t buy here. But I’d hold tight onto my shares if I were a shareholder and would increase my position on key supports.

Because man… What. A. Quarter.

As for my personal favorite,

The bottom line is crystal clear: AI delivers concrete ROI and we need more compute. That comes with more spending hence more cash generation for hardware names which remain the strongest names on the market lately.

Soitec’s Cherry on Top

Management felt generous yesterday.

Before turning to the quarter, let me briefly explain why we decided to bring forward today's communication. Since the start of the second quarter, customer demand for Photonic SOI has accelerated faster than expected. At the same time, the action plan I set out to increase our Photonic SOI production output has started to deliver positive results. These developments gave us materially greater visibility on the pace of our Photonic SOI ramp. Consistent with our commitment of transparency, rigor, and timely communication, we choose to update you ahead of schedule.

They had great news to share, so they couldn’t wait. Weird, but why not.

The thesis is playing out exactly as expected. While everyone fights to be involved in the next “winner” hardware, I believe the higher you are on the supply chain, the better. You can hardly be higher than Soitec, the company selling the wafer for any optical scale-up, from pluggable to XPOs.

Most of the revenue we see right now, and most of our revenue for FY27 will be on pluggable and partly NPO as well, probably starting. We still see co-package optic ramping up at the end of the year on the scale-out. Especially on the switching racks on the scale-out.

Photonics-SOI technology has already demonstrated its ability to address different scale-up configurations, such as pluggable transceivers, Near-Package Optics (NPO) and Co-Packaged Optics (CPO).

As long as optics scales, Soitec will. And optics is set to scale for years.

This visibility is underpinned by accelerating adoption across AI data centers, expanding multi-year customer commitments to secure capacity with cash deposits, a strong signal of engagement to securing long-term supply with Soitec. Further volumes are under active discussion.

So, without surprises, revenues are up 23% YoY compared to a 15% guidance, with Photonics‑SOI as the main growth driver, up ~100% YoY, with Q3‑26 expected to come in at 30%+ due to photonics demand.

Since the start of Q2’27, the Group has seen further acceleration in customer demand for Photonics-SOI, as well as the satisfactory results of the action plan initiated to increase its production output.

Assuming no material disruption in the AI market, Photonics-SOI FY27 revenue is expected to more than double the slightly above $100m generated in FY26.

Numbers still looks weak sequentially, but this is because of scaling. We are still early and management hasn’t fully reprioritized its production lines, so sequential growth seems low – even if well above guidance. The acceleration will start now as the thesis is now confirmed.

**Precision: their FY26 closed last quarter, today is qualified as Q1-27 so they are indeed guiding to $200M+ photonics related revenues for the next four quarters.

The other verticals are healthy enough with a small bounce in automotive - driven by new tech in cars, and expected stability through the year, while the smartphone market remains filled with inventory for core products and healthy demand for new technologies, slowly scaling. Both should be stable at best, not perfect but good enough.

Management shared last quarter that they could rapidly and cheaply refocus their fabs to meet growing photonics demand. They confirmed today that their Singapore fab is ready for Photonics‑SOI high‑volume manufacturing with a first customer confirmed. This was supposed to be done by year‑end but they accelerated to meet aggressive demand. Both of their Photonics‑SOI fabs can be expanded to meet accelerating demand if needed. Management is clear that they will work to meet demand, which comes in high volume and at better margins.

This quarter is the confirmation of our original thesis which was based on maybes. These maybes are now confirmed fundamentally, and we’ll need to see them in the numbers over the next quarters.

The stock pushed on potential until now. Today’s confirmations should be enough to justify more push as the market prices success, not just potential, boosted by cash generation improvements. We just have to wait for this now.

As to how I intend to manage my position,

SaaS Fight Back

ServiceNow confirmed what we knew: demand for AI services is real. Just like with software, companies capable of bundling a service into an optimized and affordable workflow will have customers.

The quarter beat expectations and guidance, with interesting but mixed data. This isn’t the inflection point SaaS bulls are waiting for, but it gives us clear signs that AI isn’t going to disrupt them out of existence. I think we can safely say the bottom is in for the sector, even fundamentally.

On the numbers, revenues are accelerating at the expense of margins - part of it due to a more expensive acquisition than expected, but there’s also an impact from AI token costs.

Revenue growth is still strong considering the “AI will kill SaaS” narrative, which we should bury by now. This quarter proves that idea wrong.

RPO and current RPO – to be recognized in the next 12 months, have stable growth on a large base, which is impressive - like for Google. If AI and LLMs were replacing SaaS, this growth would already have collapsed: demand is for AI-enhanced services.

We can note though that cRPO is growing faster than total RPO; looks like enterprises are taking one‑year subscriptions to see what happens, without committing long term too soon. With 98% retention for the quarter, we can assume most of them will renew, but that depends on confirmed AI value‑added.

AI demand is clearly the source of that demand.

The number of customers with agentic AI in production has increased 9x over the last nine months, a leading indicator for the future consumption opportunity.

ServiceNow AI continued to outperform expectations in Q2, with ACV crossing over $1 billion and net new ACV growth accelerating sequentially, growing over 40% quarter-over-quarter.

Deals including five or more ServiceNow AI products grew 5.5x YoY, which drove a tripling of million-dollar-plus deals.

The percentage of renewal customers purchasing agentic AI for the first time doubled quarter-over-quarter and year-over-year. Customers that weren’t already on the AI journey are signing up fast.

Interesting quotes which show in the numbers: more high‑paying customers, proving that average contract value is increasing, which usually comes with AI services.

There’s a lot of good in this quarter. The best is that it proves AI services demand, and shows that companies aren’t using LLMs to replace SaaS, as the market once feared – sometimes still fears. This was obvious, but now we have data.

Are customers going to build their own? I've yet to meet a customer who would even consider it. The best tech leaders know it will cost 5x to 10x to build an agent versus run one on ServiceNow.

What it doesn’t prove is long‑term commitments. It proves interest, maybe curiosity, and gives signs of where demand is headed, but doesn’t confirm with long‑term acceleration. This looks like what the market needs to bottom SaaS names – and I believe they have, but not enough to start a new uptrend, even on optimism.

I could be wrong and we could see re-ratings on optimism alone, but it seems early to me, given the signals.

On the contrary, it isn’t early for cyber‑security SaaS especially after OpenAI’s LLM model attack this week. ServiceNow is integrated within cyber companies, so this could indirectly be a catalyst for the company. But for non‑cyber‑focused SaaS, this still looks like early innings.

Great signals, two quarters in a row.

The Power is Out

We had both Texas Instruments and STM earnings, both were excellent. Yet, TXN is down ~5% and STM ~17% while beating expectations, with no guidance raise.

Demand for any kind of product related to power or data‑center hardware is growing both sequentially and YoY, and both companies are talking about rising demand.

During the quarter, demand increased further with strong bookings and book-to-bill close to 2 overall. We were well above one in whole end markets and significantly above 2 in communication equipment, computer and peripheral, mostly driven by optical connectivity, including silicon photonics. We saw improved visibility and signs of tight supply in several product categories.

STMicroelectronics Q2-26 Earning Call

STM raised its guidance for data centers and expects data‑center revenues above $1B by FY26 and well above $2B by FY27, with improved margins, strong traction in photonics with multiple design wins, and strong interest in solar and battery storage.

Both calls are bullish for my investment thesis: increased photonics demand means more photonics hardware, and strong interest for BESS and analog is the thesis for data‑center expansions. But the market is fading good news, just as it has for some time now, and this isn’t really bullish… Especially after these earnings.

All these names will become great opportunities in time. For now, fundamentals are improving everywhere, but narrative remains the driver #1 of stock prices and it isn’t really optimistic right now.

Patience.

Market Overview

Everything is being sold pre‑market. Google is down ~5%, Meta ~3% and Amazon ~2.5% as well. Power semis are all down, SaaS is flat… only Soitec is still up ~20%, but I wouldn’t be surprised to see that fade if the U.S. market opens red.

I won’t try to predict the market’s reactions. But earnings were good, great in many cases – and that was already true last week with TSM AEHR ASML or SMCI, but the market doesn’t care. Tensions in the Middle East might not be helping, oil is rising again; it’s hard to know exactly what and why. What we do know is that good news aren’t bought, and until they are, we shouldn’t be too aggressive.

All these names have a long way ahead of them and are great opportunities, but their stocks still need time. Consolidation – despite great earnings, remains my base case.

Slow and steady during tough market environments.
No reason to increase risk here.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

This is something.

There’s lot of anger lately on social medias as moods follow volatility. These emotions are what make investing difficult, they force us to take rash decisions; but their real source is frustration. Large drawdowns make us doubt ourselves, our stock picking, strategy, execution… They happen because we did something wrong. What matters today is to define objectively what.

I’d also invite you to look back. We all gave back a lot of returns but most of our YTD should still be green, at levels reasonable investors would call excellent. My portfolio is still up 47% YTD and the stock‑picking portfolio is up 55.59% YTD.

We are only in July, there is strength in this market, room to find performance, and my system has already improved a lot this year. I know the next ones will be better. These are great times to learn and improve, not to cry over losses and give up. No one said this would be easy.

Now,

This market is sending tough signals. Late‑bull‑market vibes, crypto‑like swings... As I’ve said many times, we aren’t early and we should all acknowledge this.

The S&P isn’t in an awful spot. We’ve seen some distribution with rising volume on flat price action and lost supports, but we are still holding the D50 and the expansion we had in early June needed to breath a bit.

More worrying are the QQQ and SMH (the semi ETF), which both lost their D50 on higher volume; buyers weren’t strong enough to step up at those prices, and who can blame them when most analysts look at 2027 and 2028 expectations to value them… Buyers need some time there while strength is moving to other verticals.

This isn’t the only ETF taking a break. Most of the YTD winners are struggling: clean energy, power semis, space, robotics… Many are now sitting on their W50 and this feels like a make‑or‑break moment for them.

We’ve seen that good news – cold PPI & CPI plus with excellent hardware earnings, didn’t end up in green candles. That says a lot about risk appetite, and fits well with these names losing key trendlines.

But I wouldn’t be too bearish either. It does feel like a make‑or‑break moment but we’ve had lots of them over the last two years, and even if these were to break, some other sectors look much better. We’ll get into those, their leaders, what I expect from them, and the different pockets of strength and weakness.

Some notes on Kimi,

This is a DeepSeek‑like moment. K3 is the latest open‑source Chinese model which, on benchmarks, is slightly below the latest U.S. models.

Another Chinese models trained with less compute which competes with the most advanced western ones. This shouldn’t impact the market as we know since Deepseek that training can be optimized; the real narrative now is inference, which is why most of the returns have shifted from GPUs to networking as the needs are to meet massive compute demand as efficiently as possible, not train models.

This confirms that models are now raining, just like Nebius predicted, and aren’t the most important part of AI anymore, just like Karp said more than two years ago; what matters is the intelligence built on top of them. This is the current need for enterprises alongside speed and cost.

What we need today is optimized compute, energy and that layer of intelligence to meet current needs; we’re far from it but the market already anticipated parts of it, pricing success for speculative assets, which makes returns unlikely to happen where most look at.

Weakness & Strength

My focus this year was on a handful of sectors: mainly AI hardware, energy and space. Today, strength is shifting into healthcare and software, which is the logical next step of the AI trade.

But I wouldn’t throw the others away.

Space,

is the weakest sector. I exited entirely. I’m aware of the fundamentals and bull case, but as a concentrated investor I can’t hold weakness. My biggest worry is that even leaders like Rocket Lab are struggling. In this situation, the only question worth asking is.

Why didn’t buyers step up on improving fundamentals and a perfect trend retest?

There are millions of possible answers; one of them being “we are in July, investors are on holiday, we’ll go higher later”. But I have to play probabilities. So the valuable question is different, because the why rarely matters in the market, what matters is how to leverage the signals in front of you.

The data says fundamentals are improving (Iridium acquisition, accelerating launch schedules, growing revenues & more), and yet, it wasn’t enough for buyers to step up and defend a trend that hadn’t been lost in two years. So the question is.

Is it worth taking the risk of buying a broken downtrends?

This could be a fluke. I actually believe it is and that space stocks will make new highs before year‑end. But opinion is just a beautiful name for bias and I don’t invest on bias. I invest on signals, data, and probabilities.

Today, space isn’t giving the signals I need to be heavily involved. Bulls aren’t buying on key trends so I’d rather be somewhere else.

Photonics,

Photonics is showing more strength with leaders holding supports. We’ll talk about Soitec but I have other names to illustrate my point - Applied Opto, GlobalFoundries, Coherent… GlobalFoundries here as an example.

We all have different books, so we should treat signals differently. I learnt after buying space names a few weeks ago, that this kind of violent drawdown has to be treated carefully: a W50 retest can still fail and burn you. GlobalFoundries looks much better than Rocket Lab. Here you’d get a great price for an asset involved at the top of the photonics supply chain, received governmental grants from quantum, in a tight environment with massive demand and limited supply as confirmed by TSM.

If I had to play probabilities, I’d play them here, with a tight risk management.

AI Hardware names,

AI hardware names are a beast of their own. Yes, they are down a lot, but most were so extended that even these ~30% drawdowns haven’t brought them back to key supports. Many great accumulation areas on the sector, nothing else.

I have no more comment. These names are extremely strong, fundamentally and in price action.

Software,

slowly becoming my favorite hunting ground. I have a very long watchlist and most charts look like they’re in the early innings of bull trends after a massive beat‑down these last months.

The realization that models are only one part of the equation and that the intelligence and data layer built on top is the real value‑added will create winners with returns comparable to photonics names this year.

The companies that can provide real value at affordable cost will change the world.

It starts with cyber‑security and continues with data owners/companies that leverage data to provide those services, and will slowly expand to any kind of service, from B2C to more passive uses of AI.

Most cyber charts look like Qualys: the market suddenly realized AI was a tailwind and bought every cyber name, hard. There’s nothing actionable here for me, but I flagged those names weeks ago; some of you hopefully got some returns.

I’ll need to be patient to get setups I can buy – I already found one this week, but I know many will come over the next months.

In brief,

Outside of space, most sectors and names I follow are still in uptrends, with leaders still leading and bulls stepping up where they should, which is why I wouldn’t panic.

I’d still be cautious and treat retests with tight risk management. My base case from here is consolidation, exactly like we had at the end of 2025, with a bounce from here as a possibility and a bear market my less likely scenario.

I’ll illustrate my base case with Nebius.

50% drawdowns are to be expected in growth names. What matters is that bulls step up where they should, just like they did back then. W50 were bought and held. As long as bulls step up on trends, there isn’t much to do but accumulate strong assets, with strict risk management.

What you see today on Nebius specifically is comparable.

If we consolidate like we did back then and see a few W50 retests, accumulation it is. If we bounce and push higher during earnings season, then great. But if bulls do not step up on the W50, we’ll be better off somewhere else. Consolidation can go on for months without being an issue, losing supports is an issue.

Do not exclude any scenario. Have your base case, expect the worst, and know in advance how you’ll behave in each.

My Portfolio & Positions

I made a lot of changes over the last few weeks, because I made a lot of mistakes. Just like any of us. But I learn from those; mistakes I used to make two years ago weren’t repeated, so I know I won’t repeat the ones I made these last two months.

Today, I own five positions: Silicom, Soitec, Meta, SolarEdge and UiPath, at 70% margin.

On Friday, I sold Tempus, ARKG and Arista, after selling Fluence, Intuitive Machines and AST SpaceMobile earlier. I already detailed the why for the last three, but not for the first three as those were done Friday post‑market, so let me go over them.

Tempus lost its W21 on Friday. That’s weakness; small, but in today’s market it has to be punished. I didn’t do this a month ago and it cost me a lot, so I adapt and factor the environment into how I manage positions.

This isn’t a bad chart and the healthcare thesis remains strong. But I have a better thesis elsewhere and needed liquidity. This was also the reason to close ARKG: I needed cash for my latest position.

I still consider Arista one of the best AI hardware stock on the market. Fundamentals are perfect, management is perfect, narrative is perfect, chart is perfect, behavior is perfect… I don’t think you can ask for a better name.

The reason for selling was the same. As a concentrated investor, I need to concentrate. I could keep a foot in Arista and build a 30‑stock portfolio, post screenshots of my 2% positions up 100% and brag on social media. But that’s not why I’m here. I’m here to increase my net worth, and that doesn’t happen with 2% stakes in 50 great stocks. It happens with 30% stakes in one.

As I said, I’ve received a lot of criticism lately on my new positions and sales. I don’t do it because I enjoy paying transaction fees, but because I try to find the best proxy for net‑worth increase, that requires concentration, which requires decisions – including cutting clear winners to buy what I see as higher potential.

Arista is a clear winner. But have you seen UiPath?

My Positions

I already went through the fundamentals a few months ago; I had the wrong timing then but the narrative hasn’t changed since, it improved, and timing is better. I’ve rarely seen such volume on any stock, and that justifies rotating liquidity into it.

Something is happening in this stock. I have no idea what exactly, but there is a lot of buying – shares and options, and the market bought it during one of the most violent downtrend of momentum driven stocks. I’ll be in as long as supports hold.

Soitec had an interesting week, closing positive – like many photonics names, with buyers stepping up heavily on its W50.

The stock is down 50% from its yearly high and my first trim, and 17% since I refilled my shares. I’m still satisfied with how I managed this position and am being patient, as I expect a Nebius‑like consolidation. It became my smallest position not because of a lack of conviction, just drawdown and patience before buying more. Environment is tight and I’d rather see strength before increasing the position, for now the stock is being rejected at every daily average retest.

The fundamentals are unchanged and unrelated to CPO delays; SOI wafers are needed for any kind of optical connectivity and are under massive demand. It’s just a matter of liquidity allocation and patience. I’ll gladly make it my largest position again once it reclaims its daily averages, and I’d cut the position if it were to lose its support.

Silicom is somehow uncorrelated to its sector, to my great pleasure. It’s also cheap considering latest news and demand for its hardware.

There is only one possible interpretation for a stock behaving better than its sector: that buyers are stepping up in size. The market sees something and buys it, while selling the rest of its sector.

I see the same thing.

Meta chose my best‑case scenario: it gave us a breakout retest and closed its gap. From my thesis write‑up:

We had our ~$640 retest, it was bought, and now we wait to see what happens. I still consider Meta the best AI name on the market here.

Lastly,

I talked about Bitcoin and Ethereum with alerts for Founding Members and trade plans in articles. Both behaved well these last weeks and I’m still holding both. I didn’t include them in the portfolio above because they are a different pocket of liquidity – just like I don’t include my bank account as cash.

If I had to choose between crypto and stocks, I’d buy stocks. But I don’t have to. And seeing Ethereum’s chart, knowing no one is positioned… I might buy it even before some stocks.

We still need confirmation of an uptrend – clear higher highs and larger volume, we aren’t there just yet… But considering the crash on some sectors, seeing crypto hold daily averages for weeks with higher highs is interesting. I’m positioned on margin and getting paid in funding. You rarely get paid to be leveraged in crypto; the few times I’ve seen this were moments before big pushes. Invalidation happens if we lose those daily averages.

It might not push higher today, nor this week, but I’d keep a close eye on cryptos over the next weeks and months.

As usual,

I will share every transaction in the chat with Founding Members, whether purchases or sales, whether controversial or not. We’ve had many interesting conversations over the last few weeks; great advice, questions and interactions.

A lot of value added.

A few last words,

Again, the last few weeks were draining. That’s what we signed up for.

I’ve expected a bounce for a few days now and Friday might have been the first sign of it. I expect it to be just a bounce for weak names and would use to cut them, not to reinforce a “I knew it, we’re so back” bias. There’s always a bull run somewhere; there’s no point holding weakness and hoping for the least probable scenario while you can concentrate on the best ones; names that were green over the last weeks should over perform in the event of a bounce, and continue higher later on.

We are entering the heart of earnings season. I expect everything to point in the same direction: more hardware, more capex, more spending, and software slowly proving that AI is a tailwind. Nothing different fundamentally. The real signal will be how the market reacts to the news in each sector.

No selling on bad news is bullish.
Selling on good news is bearish.

Disclaimer: I am not a licensed financial advisor, analyst, or broker. This content reflects my personal opinions and investment decisions for informational and educational purposes only. I hold positions in securities discussed and may buy or sell without notice. Nothing here constitutes a recommendation to buy, sell, or hold any security. Past performance does not guarantee future results.

Always conduct your own research and consult a qualified professional before making investment decisions. I accept no responsibility for any financial losses.

Liquidity is exiting a few key sectors (semi, power semi, photonics, space and data center buildouts) toward new ones (healthcare, SaaS, consumers & the Mag7, on which I am positioned).

This is why the S&P 500 is still flat and close to its ATH while most portfolios have been smashed. Liquidity is rotating, not leaving, which is bullish for longer-timeframe trends, but harsh on those over-exposed to falling sectors.

The fundamentals remain strong; returns are just rotating.
And I wouldn't expect this to change soon…

Rotation usually lasts a few months. It’s hard to anticipate the timing; the last one lasted from ~Nov’25 to March’26. This one started early June, so we’re a good month in and it’s been particularly violent, with many names already down ~50% and close to their W50. If we look at key ETFs, the VanEck semiconductor one for example, it’s losing its D50 today, so I wouldn’t expect this correction to stop tomorrow.

It can always happen, with a large V-shaped recovery, but considering the latest news and the market’s reaction… I’ll put my money somewhere else, because I’m seeing signs I don’t like much.

Before starting,

I’d like to re-introduce some core concepts in investing. Fundamentals are one thing, price action is another. You only make money when price goes up, and price does not go up just on fundamentals - far from it.

The most important in the market is optimism - and context.

You don’t want to buy and hold the best fundamentals; you want to buy and hold a stock with growing optimism within a bearish or neutral context. Take photonics, which ripped for months and delivered massive returns. Those returns weren’t based on concrete earnings, but on optimism for their future potential. Great news kept coming, but most of them are down 50% over the last month while fundamentals are better today than they were six months ago.

Context.

If you bought the improving fundamentals, your portfolio got destroyed. Those don’t necessarily convert into returns - not if you buy blindly without taking context and price action into consideration.

That said, we can continue.

TSM, ASML, AEHR & IBM Earnings

Let’s start with IBM, which triggered a “sell-off” on SaaS names with one sentence, which can be summarized by “our clients reduced their software spending to increase hardware spending.”

In the last few weeks of June, we saw clients shift their quarterly capex spend toward servers, storage, and memory purchases to secure supply-constrained infrastructure ahead of expected price increases. This dynamic impacted client buying patterns. While we anticipated some supply chain related impact in our expectations, we did not anticipate the magnitude of the capex reprioritization. In addition, clients were distracted with rapidly-evolving, industry-wide cybersecurity concerns in the quarter.

The stock was punished with a -25%, in line with the market’s recent violence.

The three others delivered a masterclass. They all sit at the top of the AI hardware supply chain: ASML with its lithography hardware, TSM as the largest and most advanced semi fab, and AEHR one of the leaders in semi testing.

All of them beat the market’s expectations, and the bottom line is the same for all: companies need much, much more hardware and are ready to pay the price for it. You can read detailed reviews, take accounting classes, go point by point or whatever; the conclusion will always be the same.

More hardware. No clue when demand will slow.

Strong end market demand this year has motivated our customers to aggressively add capacity on their leading-edge nodes.

Customer demand remains very strong, with visibility now extending several years into the future.

ASML Earning Call

Given the continued strong structural demand from our customers, including the newly emerging agentic AI market, we have decided to raise our full year 2026 capital budget.

If you asking about the AIs CAGR, let me give you not a number, but its stronger and stronger. We dont give you the number today because it continue to increase, so we dont know how to answer this question. Stronger than what we said before.

TSM Earning Call

So we had four tech earnings, and the overall conclusion was that their clients needed more hardware and were slowing down on software. Their cash was rotating. And yet, if we look at the market, what happened this week?

Sofware stable/up.
Hardware down.

Because context matters. Once again, we only make money if our stocks go up, not just by looking at fundamentals. The market isn’t fading AI hardware or disbelieving those earnings and comments; we simply don’t have buyers left at these prices, and as price falls, holders turn into sellers while buyers are too scared to step up early. They’ll step up - just not yet. Doom spiral for now.

When bad news triggers no reaction, you're close to a bottom.
When good news triggers selling, you shouldn't push your luck.

From here,

There’s no saying what will happen, but as usual we have to play probabilities, which aren’t favorable for a return of the uptrend in weak sectors. Those could bounce, but a real uptrend? I don’t believe it, yet.

So I’m doing what the market is telling me to do: selling weakness, buying strength.

I closed four positions today, names I still believe in but that showed weakness. And I put it all into one stock. A SaaS company I’ve followed for a while now. Fundamentals are stronger than when I first found it, valuation is much lower, and price action looks like textbook accumulation with extremely bullish option flow.

It has everything I look for, and I’ll go through the full breakdown: what I sold and why, the name I bought, why I’m building it into one of my largest positions, and what will make me increase it even more.

Read more