TL;DR - Demand and fundamentals aren’t the problem. They’re as strong as they’ve ever been. The market is pricing one thing only: financing risk. If the money keeps flowing, nothing changes, but the market believes money will stop flowing. This is about perception for now, not fundamentals. Until perception clears - either way, I’d rather sit on cash than sit on a drawdown. Better to be out than anxious to get out.
That’s the whole piece. Everything below is the why.
Many AI names on my watchlist broke south and I don’t see any short-term catalyst to turn them. Or rather, I see what could, I just don’t see it happening. This feels like space back in June, it broke my portfolio then and still hasn’t recovered. I won’t let that happen twice. My system is doing its job. Price broke, I cut. Cash is a position and it gives me the freedom to swing aggressively the moment the trend turns.
We’ll go over everything today; the two things breaking the market - wars and AI funding questions, their link, the investing logic behind my decisions, what I’m doing with my book, and some other thoughts I want to share.
Top down.
Wars, oil & stupidity
We have two wars going on. Everyone knows that. Russia and Ukraine have been at it for years and Iran and the U.S. are starting to look just as stupid. The market had learned to live with both, but not their recent escalations, as both hit the same nerve: oil - still the heart of the world’s economy.
Ukraine’s gone rogue on Russian oil infrastructure. Russia was the largest and last “safe” source of production, Trump even struck a diesel deal with them. It should tell you something that Trump is cutting deals with Putine in today’s geopolitics. It says the U.S. needed it badly. Yet, Zelensky keeps hitting refineries, openly ignoring Trump. This isn’t a very good sign.
In the Middle East, Iran struck two UAE tankers in the Strait of Hormuz, the Houthis attacked Saudi oil infrastructure, ships continue to be targeted and Iran said it would intensify. The strait is open, yes, but with the added cost of logistics and insurance, prices aren’t getting any relief. Plus damages will need repair so even a fix today will need months to be felt in the economy.
Everyone seems to agree on one - terrible, course of action: use oil as leverage. They’re holding the world hostage to their cause and are pushing everyone to their limit, playing to see who breaks first.
And it’s working.
Inflation is back - energy-driven, but energy bleeds into everything, so today’s energy spike is tomorrow’s pricier services and products if left unchecked. The market knows it. That’s why the US10Y has been climbing for weeks and a rising 10Y hurts every American, and every business that needs to borrow.
You know who borrows a lot lately? AI.
Yes, oil impacts AI
Not directly. But it does.
Yield and returns work together. People cry about CoreWeave’s yield but it wouldn’t matter in a normal economy, compared to the returns CoreWeave could earn on money borrowed earlier ~9%. Those yields aren’t the heart of the issue.
The issue is that CoreWeave - and AI names, need constant investment to return above their contracted yield. They need acceleration. That means borrowing more, now, while cash generation today sits below the yield on existing debt, and well below the yield on any new debt, even at preferential rates.
So the market’s question is: who funds this?
Nobody questions the fundamentals or AI’s potential. But fundamentals are only one part of what moves a stock, and not the biggest. Narrative and liquidity come first, I’ve said that a lot. Everybody questions whether AI can keep receiving funds.
And the market is clearly telling us what it wants to lend to CoreWeave - and others. And that yield is way too much for CoreWeave, and less would be way too risky for lenders. Why take the risk? The only players who could are AI companies - Broadcom and Nvidia for example, and they’d be right to do so as their business relies on AI funding until compute providers and frontier models generate cash. BlackRock or other funds? They don’t.
Today’s question isn’t only about fundamentals. It’s about whether the money comes and where it comes from to keep the wheel turning.
But it’s also about fundamentals.
Anthropic, meh
Anthropic’s quarterly numbers landed yesterday as they went to the banks for more funding, numbers in hand to justify it. Let’s see if they get it and how much, for now we can comment on the numbers.
Revenue up ~45% QoQ. Management said they were profitable “on an adjusted basis”, ran at 11% operating margin, were cash-flow positive, and generated $1.70 of revenue per dollar of compute. Great numbers, on paper.
Off paper, “adjusted basis” means everything and nothing. I’m a billionaire adjusting the “.” basis in my bank account. Without knowing what’s being adjusted, it’s hard to trust cash generation from a company that was losing $2 for every dollar it made two quarters ago. But, it could be. That would be very bullish.
And I am bullish AI.
I can comment on ARR as well which many call flattish, maybe reaching a plateau. I don’t believe it to be true as intelligence demand rotates between models, so when OAI releases a better coding model, demand rotates to them. Competition.
I personally remain bullish on these charts, and most data confirms my bullishness, including those results which are honestly good. Below market’s expectations though, which is a big problem. But there’s another one.
Besides the “adjusted” asterisks, Anthropic said it wants to invest over $500B in the next decade.
Let me illustrate. My Substack has a ~$40k ARR. Say, I spend ~$35k of it on tools and research, whatever I need to run it. Do you think I could go to the bank and have a $400,000 loan contract for 10 years at ~7% interest rates? I mean, I’d buy myself a cool house, it’s enough justification isn’t it?
Of course, I’m exaggerating. My house wouldn’t be productive like Anthropic’s models are. And I’m just a random dude not a company. But still. The market isn’t asking “is there demand?” It’s asking “how the f*ck are you going to pay for it?” and “is your growth really enough to justify it?"
Funding is everything
This is the market’s only real worry, its biggest. The market isn’t being stupid here. It’s pricing risk, with reason.
Rates are high, AI labs and compute providers are already heavily leveraged with no to low free cash flow. They could have cash flow, they are cash machines after all, but only by stopping investment - which they won’t and which would wreck the market anyway. AI only works while liquidity flows, and liquidity might be drying up. Two years ago, OpenAI could say they needed $500B. We had cash reserves and low yield, and the market believed it. Conditions are different now.
Perception.
Nobody has the full picture on funding or the terms being negotiated behind closed doors; not retail, not even the big players. This is what the AI trade is today, this is what you’d be buying. Not fundamentals. Not narratives.
But whether capital keeps flowing or not.
Right now the market is betting it won’t, that we’re hitting the ceiling of what today’s conditions can finance. I don’t know if it’s right or wrong. I’m just a retail. I’m just aware of what buying AI stocks today means. Forget fundamentals.
What could help?
Lower rates.
How do we get those?
The Fed’s answer is hikes, which is stupid, because this isn’t a monetary problem, it’s a geopolitical one. The only source of inflation here is oil, due to the wars. Hiking into that just adds onto U.S. debt which will only raise the US10Y more, and squeeze the AI trade by doing so. That’s stupid but they still did it.
No, only stopping the wars breaks the loop.
higher oil prices → inflation → higher rates → less funding → AI wheel slows/stops.
Given where things stand - Zelensky off the leash, Iran knowing exactly where to push before the mid-terms and Trump refusing to sign a bad deal to save face… The question is who breaks first? And when? It could be no one, before long.
This is also what you’d be buying with AI names: betting on those conflicts’ rapid resolution.
Investing in all that noise
Now the semantics.
The FinX comments are all the same. “Buy the FUD” “Don’t be a pussy” “Big-boy pants on and buy” Or the famous and misquoted Buffett line.
Stupid, all of them.
I believe in AI as much as anyone: insatiable demand that will generate enormous wealth for many. But to make money, someone has to buy your stock higher, which requires optimism; for optimism, and to build anything, you need funding.
Buying today isn’t about fundamentals. Again. You are buying funding capacities and geopolitical decisions. If both resolve positively, you’ll make money. If they don’t, you’ll lose. But ASML reporting massive demand or TSM talking about spending and visibility for two years today won’t change anything.
if you expect Trump to come to save the market like he did with tariffs - and as I read on social medias, it could happen, but to save the market. Not growth stocks retail gamblers. The S&P500 is still near all time high, nothing to do with the ~20% drawdown post tariff news.
This is my view of the situation today. And my problem is that buying here is about hope, not thesis. You don’t size a position on faith that the right person blinks first.
I would also go a bit further on social medias and echo chambers.
If you are a bull, everyone on X will tell you how cheap everything is, how it’s the opportunity of a lifetime, because that’s how social media works. That’s how we get paid: reinforcing bias. That’s how communities get built. That’s how content is spread: echo chambers.
I am not saying most do it on purpose with bad intentions. Most just do what they feel like doing. What I am saying is that they might be right. They might be wrong. Just know that the day the music stops, you’ll be reading the same post about cheap valuations and once-in-a-lifetime setups.
Except this time the market won’t bounce back.
Social media isn’t real life, and it isn’t the market. We - the retail crowd posting all day, are fishes. The funds that actually set prices don’t manage risk the way we do, and right now they’re the ones selling.
I know this is a long write-up. Take a breather, like the market.
(too soon, sorry).
Sometimes, I need lots of words.
There are always new learnings
This is important here, because it will impact what I’ll do moving forward.
My whole point with this Substack is to learn how to make money and share while I do. The system is meant to improve with every mistake. I’ve made plenty and I’ll make twice as many - hopefully. That’s fine.
I am a big fan of Anand and believe he is one of the best trend analysts on X, and for free. He saw both corrections coming this year, and has the perfect mindset, so let me share some of his wisdom.
Digest this. It’ll help.
What’s not fine is making a mistake and not applying the learnings. I’d be no better than a child. I got burnt in June by the space sector, expecting a bounce, hoping for some news. I wouldn’t be any better if I were to reproduce the same mistake here, despites my own convictions - which you know are extremely strong on both names.
The last weeks also taught me plenty.
I shared the main one already: take profits. I’m a swing trader at heart, but I built my system around the home runs I used to miss - Palantir and Nebius to name two, great positions I stopped riding because of bad habits. I fixed those to land home runs. Real money, not peanuts.
My mistake was thinking I had to change everything, all the time. Some conditions are good for home runs. Others are for taking profits and de-risking. Not because the assets are different but because the conditions are.
I can be both. I actually have to be both.
This is not a home-run environment. Not a buy-and-hold one. And definitely not a buy-every-dip one. I don’t live on social media except to write, but I follow a handful of investors closely - Anand included. Most shortened their timeframes and made a killing.
My own trades these last months were excellent; Tenable, Tempus, AXTI, UiPath, Meta, Monolithic… Clean, well-weighted entries. This is what I do extremely well now. I know when to buy, why and where - at least I have done it well. The issues were the exits, because I ignored the environment.
That changes.
I already know how: keep doing what I’m good at, fix what I did wrong.
Le cul propre
Before concluding with what I’m doing, let me tell you a story - it matters.
We’re back late 2021. The crypto run was still going, back to ATH after a massive - but usual, 30% drawdown. We didn’t yet know it was just a dead-cat bounce. I was young, excited, making “a lot” of money.
I followed a known French investor. Like many. And he started talking about selling everything. He called it “cul propre”, which means “clean ass” in French, because, as he put it, a clean ass can always get dirty again.
The logic was tight: sell, book profits, wait. If the market pushed higher, it was volatile enough that he’d still catch a short-term trade. He wouldn’t get the full move but who cares? Being all in cash buys you the one thing that matters most in a crazy market: the freedom to hit a great setup with a massive position, maybe even an all-in.
He could swing. He just wouldn’t hold. Better out and ready to be aggressive than in and desperate to get out. That’s what he said.
I laughed at him. Many did, in between insults - I didn’t insult though don’t do that. Who the f*ck sells at all-time highs when crypto is the next big thing? What an idiot, this self-made multimillionaire. Turns out he swung an altcoin with massive size and double digit gains. Then shorted the market.
What an idiot. Lol.
The moral isn’t “sell everything”. It is that you can always buy back and make a fortune on a shorter-timeframe trade, without the risk of holding dangerous conditions. You won’t die because you sold. You might get wrecked by not selling.
In the next weeks, I gave back everything I made during the bull run.
He walked away with a few more million.
