A month ago, I sold my AI stocks.
Did I stop believing in AI? In how it’ll reshape society and the wealth it’ll create over the years? In the buildouts, the hardware, the energy it needs? Do I hate Nebius, Axti or Soitec?
No. To every question.
But since that decision: Nebius down 21%, Applied Opto 44%, Axti 25%, Soitec 21%, Silicom 31%. Those are the main ones. Others fell in the same range.
The market didn’t care about my opinion. It shouldn’t.
I also lost ~25 paid subs and a lot more free ones. Most left politely. Some left calling me names, spamming the chats, calling me a clown unfit to write with some more insults left, right and in private. I’m not telling you this for sympathy. I’m telling you because it matters.
You have every right to unsubscribe or dislike my content, but it isn’t what happened. What happened was that I shared an opinion they disagreed with, without even trying to understand or question its validity.
If you don’t tell me what I want to hear, I don’t want to hear it.
Investing wise, that opinion turned out to be right.
Of course, I don’t have all the answers. I’m a normal guy who decided to spend most of his waking hours learning how to make money in markets because I think there’s no more useful skill in life. But you have to want to learn, and most people don’t.
This is a post-mortem on a violent October, a reminder of what the market actually is, how it works, why risk management and price action matter, with concrete examples.
Lessons I believe timeless and valuable, even though no one wants to hear them.
Markets are a liquidity contest
Argue fundamentals, valuations, multiples, moats, whatever you like. At the margin, a stock rises because there’s more buying pressure than selling pressure. That’s it.
Just like in life, we are not equal in that contest. It isn’t a fair battle of opinions, it is a battle of liquidity, and the side with more of it wins.
Let me paint an example. An innocent example.
Everyone believes Nebius is a wonderful company and its stock should go higher because their execution is great and so its stock should be.
Say a million people each hold $10k of Nebius, convinced it goes higher. That’s $10B. The same as one fund with a $10B stake. On paper, a tie. In reality, it isn’t close. The million never act as one: they’re scattered, emotional, and the second the stock drops they sell into each other and hand the move away. The fund acts in a single decision, and it doesn’t even need its cash. It can short, use margin, run options, stack leverage. One coordinated, leveraged balance sheet flattens a million uncoordinated ones every time. And the more liquidity you have, the more you can multiply it, which is exactly why the big players are the ones to set the price.
If that big fund sells, the stock goes down. Make no mistake, that big fund not only has more money than all of us but it also manages it differently, its decisions aren’t based on the same framework ours are; its decisions can look stupid to you, but they aren’t. You play a different game on the same board, but you play it with his rules.
That is what the market is.
Stocks go up for a reason
Liquidity moves price. But liquidity itself moves on something, the big players still buy, sell and short on a framework.
They’re smarter, better informed, better resourced, and they outwork you. The only thing you share with them is that you both make decisions; theirs simply weigh more and account for more than yours.
Here’s where almost everyone on FinX is wrong - dangerously wrong, as we saw this month. Everyone studies fundamentals and makes assumptions.
No one studies why a stock goes up.
Thinking fundamentals move stocks is backwards. Expectations move stocks. The two are not the same. Fundamentals can be flawless while expectations rot, and it doesn’t even matter whether those expectations turn out true, because as long as enough liquidity believes them - or their risk, price follows.
Ignoring this is how you go broke
Every stock has a reason to rise. It’s rooted in fundamentals, yes, but there’s always an underlying expectation driving it, and our only job as stock pickers is to find that expectation and respect its changes - or respect the people who do.
Let’s take another innocent example.
Why did Nebius go from $20 to $290 in a year? AI demand? Compute quality? The Microsoft deal? The CEO’s hair? None of it.
If it were those things, CoreWeave and Oracle would’ve done the same. They didn’t. CoreWeave ran, but erratically and nowhere near Nebius’s ~14x. Oracle had one good stretch, then the market dumped it. Same service, comparable execution, wildly different stocks. If this were purely compute and fundamentals, they’d have risen together healthily as demand and pricing improved. They didn’t.
Nebius was a winner because of its balance sheet.
More precisely, what the balance sheet let it do: accelerate faster than anyone else. The only name that matched Nebius in the market was Iren, for the same reason: land and power already in place, which let it accelerate faster too. So it wasn’t even the balance sheet. It was acceleration potential. That, and only that, is what the market bought from $20 to $290 on Nebius and $10 to $73 on Iren.
Spot that, and you don’t just buy early, you also know when to leave. Not when fundamentals crack - they still haven’t, but when the signs that what the market is buying might crack start showing. Nebius’ balance sheet was my early investment thesis, based on its fundamentals.
These assets allowed the company to start their business easily, with a datacenter ready to be used, cash to expand and equities/subsidies allowing them to fund themselves more easily, through the market - which they did already, or through classic debt contraction - which they didn’t yet.
First Nebius Deep Dive, Jan 27, 2025
Nebius holds $2.3B of cash today without any debt, cash which, as said, is used for growing and operating their unprofitable business. This amount would easily allow them to run unprofitably for years, even after investing the expected $1B on their infrastructure.
Fundamentals themselves weren’t the story. Acceleration capacity was.
This is where ego shows up. When I shared this weeks ago, I got called names. When I said this week that Nebius could follow AST SpaceMobile’s pattern, I got called more. Go read the replies if you want a masterclass in denial.
Why do I think this? Because regardless of sector or fundamentals, both were bought for the same reason: acceleration and ROI expectations after a capex build-out.
The truth is often simple: if Nebius, like AST SpaceMobile, stops accelerating for any reason - including ones outside its control, the market won’t reward it. Best case it stalls. What actually happened is it got sold, hard.
Disagree if you want. Bring arguments. I’ll refuse “it’s irrational” and “it’s manipulation” - those aren’t arguments, they’re toddlers’ tantrums.
Timeframe & selection matter
I’ll mitigate my point to be honest, and add context.
I disagree with most on social media, but I respect a few - M.V., StockMarketNerd and Type-F (Emir) among them. They invest purely on fundamentals and (intrinsic) value, something I’d never do because I don’t think I’m smart enough to. They might be, the numbers show it, so they’re clearly doing something right. But their edge on me is that they can do something I can’t: hold through brutal drawdowns, because their edge is a long timeframe and a read on fundamentals I don’t trust myself to have.
I believe, like M.V., that Nebius will be worth more in 2030 than today. But holding that means potentially sitting through months or years of pain for a result that could still be wrong - not every pick wins, and I won’t marry a name for years on an opinion only to be wrong at the end.
Look at Cisco: two decades to make new highs. I’m a network engineer. Cisco’s fundamentals were never weak in that stretch. The market just saw no reason to pay up for a company doing exactly what it should.
Healthy operations don’t guarantee returns. Ask PayPal holders.
And timeframe is only half of it, selection matters too. There’s a world of difference between holding quality, established companies for years and holding growth names. Don’t quote Buffett to me if your book is three stocks and they’re Nebius, Axti and Iren. That’s brain dead. Quality compounders can be held for years with some peace of mind. Growth names? The volatility and the risk are multiplied, and for every Palantir that works there are twenty Carvanas, Pelotons and Zooms that don’t.
Want to hold for a decade? Fine, but you first have to be right, which is brutally hard, and in the meantime, over the next two years or so, expectations and liquidity rule while fundamentals wobble and try to shake you out.
No success story runs in a straight line. You’ll have to stomach the bumps.
Price action works
It does. It’s just misunderstood.
It isn’t an exact science, it’s a read on what buyers and sellers think of price. Of what liquidity does where. Today, humans managing liquidity are the minority. Algorithms do most of the work, and algorithms are hard-coded: they don’t think, they react.
So explain why maximum pain hit right as stocks lost their W50; not one stock, every AI name’s drawdown accelerated below it. Explain why W50s and other hard supports, once broken, got sold on every retest in a downtrend and bounced on every retest in an uptrend. It’s not mystical: long-term trend levels are where the biggest liquidity pools place their buys and their stops, so reactions at those levels are information.
If you do not know why your stock moves up or down, at least respect the information liquidity is giving you for free, because when buyers do not step up where they should, it means big liquidity doesn’t want to or cannot, for whichever reason.
That doesn’t make price action a crystal ball. It makes it a dataset you’d be stupid to ignore, even as a long-term investor, if only to manage risk and buy at better prices.
With all due respect.
So what?
I can’t hand you the perfect recipe yet. That’s what I’m building through this Substack, and I believe I’ll get there. It’ll never be perfect - nothing is in the markets, but it’ll be good enough to mechanically outperform without the headaches.
It has been so far, and improving healthily, avoiding the most violent part of this drawdown being a proof of it. The first part of the story is building the system. The second will be to put it to work and automate it, which is where we’re going next.
It’ll run on concrete rules, not bias.
Always know what moves a stock.
Respect price action.
Acknowledge your limitations.
Protect your capital.
Last Thoughts
Investing is hard. It’ll crush you.
But investing without wanting to learn, without at least testing the basic truths before dismissing them, is how you destroy yourself. You can believe the Earth is flat, that the sun orbits us, that nuclear weapons are a hoax.
That’ll make you a happy fool without wrecking your life.
Refusing basic markets facts - like that people far smarter than you have tried to hold growth names on fundamentals alone with no risk management and been destroyed, won’t leave you a happy fool. It’ll take your net worth, your capital, and eventually your will to keep going.
We’ve seen many through the years. Many more this week losing 6 to 7 digits. You just have to open Twitter to see it.
For anyone who thinks I’m preaching: look around. Over the last two to three years, while going through a long and strong bull market, who actually made money, kept it, and was transparent about it?
The ones who respected the market for what it is, a liquidity exchange. People like Anand, Bulltard, SixSigma.
The ones who didn’t chain themselves to AI growth names and bottlenecks.
Most of the rest got run over this month, despite strong convictions which might end up being right in a few years, but liquidated them or left them with no capacity to buy more of their favorite stocks today, sitting on a -50% book.
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.