AI stock correction 2025: debt, the Fed and sector rotation
Felix Nikolas Prehn explains why AI stocks have fallen 40 to 60 per cent from their highs and where the opportunities may lie.
Felix Nikolas Prehn, economist and former investment banker
Listen on YouTube
The AI stock correction of 2025 has wiped hundreds of billions of dollars from companies such as Oracle, IonQ and Bloom Energy, even as the S&P 500 has risen nearly 18 per cent. Felix Nikolas Prehn, an economist and former investment banker, traces the sell-off to three interconnected factors: heavy debt financing by AI infrastructure firms whose revenues remain years away, the Federal Reserve's precarious position with real interest rates barely positive and inflation still above target, and a broad rotation from growth stocks into value stocks such as utilities, defence and healthcare. He draws parallels with the dot-com crash and the 2008 financial crisis, noting that the underlying technology is real but valuations ran far ahead of profitability. The episode sets out a framework for evaluating holdings, warns against averaging down on fundamentally weak companies, and highlights picks-and-shovels plays in power, cooling and semiconductor equipment.
In this episode
- Scale of AI stock losses from 52-week highs
- Historical parallels with dot-com and 2008 crashes
- Debt-driven financing behind AI infrastructure spending
- The Fed's rate cuts and the inflation trap
- Growth to value rotation explained
- Action plan for holders of beaten-down AI stocks
- Mistakes to avoid including revenge trading and averaging down
- Picks-and-shovels opportunities in power and biotech
Transcript
If you own any AI stocks right now, what I'm about to show you could be the difference between protecting your wealth and watching it evaporate going into 2026. The AI stock correction has already wiped out literally hundreds of billions of dollars in market value. We've got Oracle down 48%. We've got Olo down 61%. INQ down 46%. And this isn't just about a few stocks. This is about a big wealth transfer happening right now while most investors are still completely unprepared. The question isn't if you'll be affected, it's whether you'll be on the winning side or on the losing side of what is actually coming next.
My name is Felix P. I'm an ex investment banker. That's Winston back there who's the brains behind the whole thing in the research here. And I've seen how banks work from the inside. I've been investing for over a decade and I'm also the founder of the GOAT Academy where we've taught about 20,000 students how to navigate the market more like professionals. I'm also the co-founder of TradeVision.io where we give you institutional grade news and data.
I'm dedicating my retirement to teaching you how to protect yourself from the schemes and the financial engineering out there because the truth is they're not really trying to help you, right? They're trying to get rich off you. That's kind of how it works. So what have we done for this video here? Well, Winston's analysed Fed policy, tracked every major AI stock correction, studied historical tech crashes, dot-com all the way through to today, fact checked every single number you're about to see.
This is not fear-mongering. I'm not saying AI is over. I'm not saying sell your AI stocks. This isn't clickbait. This is just facts. And by the end of the video, you will know how to position yourselves better, not just to survive this correction, but potentially profit from what's coming next.
So I'm going to break down 3 things for you. 1, the brutal numbers. Exactly what's happened to AI and momentum stocks as I'm recording this and why this correction was inevitable. 2, the hidden mechanism driving this crash. It's not what people think. It goes way deeper than overvaluation. And then 3, your exact action plan. What to actually potentially want to buy, what to avoid, how to position your portfolio for the opportunities that always emerge after these corrections. So you'll understand exactly what's happening, why it's happening, and most importantly, what you might want to do about it.
So let's dive in, shall we? Let me show you what's actually happened here. This is recording in towards December 2025. This is wealth getting destroyed, right? Brutal corrections from 52-week highs. Oracle down 48%. We've got Bloom Energy down 48%. Oracle down 61%. Centrus Energy down 52%. Iris Energy down 56%. Nvidia down 19%. These are the champions of the AI world.
Now, if your stomach just dropped because you own some of these stocks, don't panic yet. We're going to talk about what to do, but first you need to understand the facts. There is more. By the way, I didn't want to upset you too much, but there's a second slide with more carnage, but this is important. At the same time, the S&P is up almost 18%. The NASDAQ is down 12% from its recent highs. So this is not a market crash. This is a sector rotation. This is what I call selective correction. And that's actually good news if you know how to play it.
I always say follow the money. The money moves from industry to industry and that's what it's doing right now. The Fed just cut interest rates third time in 2025. Inflation is sitting at 2.9%. That's important to put that all together. So the problem is this. All of the money that flooded into AI stocks in 2023, 2024 was based on hope and hype and future potential. Now we're at the end of 2025 and the market is asking where are my profits? And a lot of these companies don't have a good answer, right? They just keep saying, "Well, AI will fix it," but they're not very specific about it.
We've seen this pattern before. So let me take you back through history briefly because if you understand the pattern, you can benefit from it. The dot-com bubble in 2000, everybody thought the internet was going to change everything. They were right about that, but they were wrong about the timing and which companies would win. Companies with zero revenue were valued in the billions just because they had a dot-com in their name. And literally my office building, it was named dot-com house. I kid you not. They've lately renamed that.
But there are companies like Pets.com. They spent $300 million on advertising, but they made basically no money. They went bankrupt. The NASDAQ dropped 78% from its peak. Companies like Amazon dropped 90%. But here's the thing, Amazon survived. If you bought Amazon in 2001, well, you'd be a squillionaire by now. So the lesson, the tech was real, the opportunity was real, but the valuations got way ahead of reality.
If you look at 2008, the problem wasn't tech, it was debt. Banks had leveraged 30 to 1, 40 to 1, and there was fraud in the housing sector. So when housing prices stopped going up the whole system collapsed. The S&P dropped 57%. But if you had the courage to buy stocks in March 2009 you would have made 400% by 2020.
What about Covid? The market drops 34%, right? Got a per cent a day. Fastest crash in history. Everyone's freaking out. The economy is done. We're going into depression, right? Hopefully pharma will save us. So what happened? Well, the Fed printed trillions of dollars. Stocks recovered. If you bought the dip, well, you would have made massive gains.
So what's happening today? We've got a sector specific correction. AI stocks ran up hard in 2023 and 2024. Nvidia went from $120 to $210 in less than 2 years. And the same story with things like INQ, 300% up and so on. But the reality is setting in. AI infrastructure costs are massive. Commercial revenues are developing. Competition is intensifying. And valuations, well, they got a little bit ahead of what's actually happening.
Now, does this mean AI is dead? No. AI is going to change the world just like the internet did. But just like 2000, we're seeing a correction where the market separates the winners from the losers. And the smart investors, well, they're positioning themselves right now to buy the winners at discount prices. At least that's my humble opinion.
Now, before I show you the real reason this correction is happening, and I think it's not what you think, I want to invite you to something special. I'm going to hold a live session between the holidays, December 27th, and I'm going to reveal why I am putting $1 million of my own money into an experiment to beat the S&P 500 in just 15 minutes a week. It's not complicated day trading. It's not spending hours glued to charts. Just the exact rule-based system that we've been working on that is based on what Wall Street does and we call a trend monster. That's the core part of that.
This is going to be inside of the private community that we have. But I'm going to walk you through how that works. And if you're curious why I would put $1 million of my own cash into something like this and aim to prove that it works, and I might well be wrong on that, by the way, I could lose that money, then join me and start your 2026 with a better understanding of how institutional frameworks attack markets like these and any other markets. There's a link down below to grab yourself a free seat for that live training. felixfriends.org/training. It's free. All you got to do is show up.
Will there be a replay? No. Why not? People don't watch them. And I know you always think, "Oh, but I'd be different. I'd watch it." Yeah, but statistically that's just not true. So we don't do replays. People just don't show up for it. If you want to show up for yourself, show up live December 27th and we're going to have some tremendous fun together.
Now, back to the hidden mechanism here that's driving this AI correction. Most people just think, oh, it's just overvalued. And yeah, there's stocks like INQ that were valued at 200 times sales. Oracle's PE was kind of mad. But that isn't actually the full story because if that was just all about valuations, all expensive stocks would be dropping. But they're not. Some are actually going up.
The real reason this correction is happening comes down to 3 interconnected factors that Wall Street doesn't really want you to understand because when you understand them, you stop being their mark, right? You start being their competition, which is really what we're doing here.
So what is it all about? Well, the first factor blew my mind when I started digging into it. These AI companies aren't just raising money through stock offerings. They're using really complex financing arrangements that involve massive amounts of debt. Oracle was in talks with Blue Owl Capital for a $10 billion funding deal to build a Michigan data centre. Apparently that deal got stalled. The stock collapsed.
Why does that matter? Because Oracle's business model for AI involves building massive data centres that cost billions and billions of dollars. But the revenue from those data centres won't come for years. Or we have companies like CoreWeave. It's a major AI infrastructure player. They have zero profit but billions in debt. Yet they secured huge partnerships with tech giants through something called equity for future profits arrangements.
So here's what's happening. These companies are promising future AI revenues to justify today's spending. But we've seen this movie before.
Right. In 2008, banks promised future housing profits to justify present-day lending. Those profits didn't materialise. Well, everything collapsed. I'm not saying we're headed for a 2008 star crisis. By the way, AI is real. The demand is real. But the timing of when these companies become profitable, that's where the risk is. And right now the market is starting to wake up to that risk.
The second factor, we're going to talk about the Fed because what they're doing right now is a very, very dangerous game. The Fed just cut interest rates down to 3.5%. Third consecutive cut in a year. On the surface that sounds great, right? It's great for stocks. Lower rates means more money flowing into the market. But the catch is that inflation is still, what is it, 2.7 or something like that, well above their 2% target.
So let's think about this. 3.5% interest rate. Say it's a 2.9% inflation. I think it's 2.7. It doesn't really matter. So what do you get to? Well, you get to about a 0.8% real interest rate. Barely positive. The United States has an almost 0% real interest rate. It's not a banana republic. It is the United States.
So the Fed's in a trap. If they cut rates too much, what do you think is going to happen to inflation? Well, inflation's going to spike, isn't it? Which, what happens if inflation spikes? Well, the Fed has to raise rates. That would crash the market. And then think about all the debt these AI companies are piling up. Well, the cost of that debt could therefore increase. So this is a high-risk strategy the Fed's playing there.
And AI companies will be the first to get hurt if the Fed is wrong. Because when money is cheap and plentiful, investors take risk. They buy speculative stocks. They bet on the future. But when money gets slightly more expensive or when there's uncertainty out there, those speculative bets are the first to get sold. People put into something more reliable. And that's what we're seeing here.
All right. Here's the third factor. And this is actually good news. In investing, there are generally two types of stocks. Growth stocks, these are companies that are growing fast but might not be profitable yet. Think IonQ, Oklo, AI startups. You buy them for what they will be worth in the future. And then you have value stocks. These are companies that are profitable now. They pay dividends. They're more stable. Utilities, banks, healthcare, that sort of thing.
What's happening right now is a rotation. Money is moving out of the growth stocks and into the value stocks. Why? Because when there is uncertainty, when interest rates are choppy, when there is geopolitical madness happening, those investors want safety. Now, what's safer? A company that's profitable today or one that's promising you it's going to be profitable in 5 years, fugazi stuff.
So utility stocks up, consumer staple stocks up, defence contractors up, healthcare is looking good, right? Gold and silver is looking extraordinary. But AI stocks are down 40 to 60%. That is actually normal market behaviour. It happens in every cycle. Growth stocks lead when times are good and everyone's optimistic. Value stocks lag and then they lead in uncertain times.
And what does it tell us? This correction in AI stocks is actually part of a healthy market cycle. That is the market saying, okay, we got a little bit too excited, right? Let's take a breather and reassess. We took a little snort of a little bit too much AI stuff. Now it hurts if you're holding these stocks, yeah. But does it mean the opportunity is over? No.
So now you understand what's happening, right? You understand the debt trap, the Fed tightrope, the growth to value rotation. But understanding does not make you money. Action makes you money. So I'm going to give you the exact playbook for navigating this correction. Now, this is of course my opinion. I'm not giving you financial advice. So be smart about it. Use the grey matter between the ears. What to do if you're already in these stocks, what to avoid, what to look for, and how to position yourself for the opportunity that comes after this.
So I'm going to give you a simple framework. Okay, it's action plan time. So what do you do if you're already holding Oracle, Nvidia, IonQ, or any of these beaten down AI stocks? Well, don't panic, right? Let me explain. When a stock drops 40%, 50% or 60%, your instinct is to stop the bleeding and sell. Well, you probably should have had a stop and got out earlier and actually kept some of those gains, but now it's too late for that. What happens if you panic sell? Well, you lock in your losses at the worst possible price probably.
So let's say you are sitting at 50% losses. Well, I'm not saying you should hold and hope. That's how people really lose money. So what have you got to do? Well, you got to evaluate the fundamentals. You got to ask yourself these questions. Is the company profitable or is it on a clear path to profitability, right?
And then you got to look at some of those companies and you say, okay, well, Nvidia, is it profitable? Are they printing money? Yes, right. What about IonQ? No. They're pre-revenue. They're burning cash. They have strong partnerships or customer commitments. Well, some, right. What about Oracle? Yeah, they got major cloud customers. TikTok, right? They're now the backbone of TikTok. Oklo has government contracts but no commercial licence yet.
So can they survive? Check how much cash they have. Check how much cash they're burning, right? Who can survive? That's where you want to be. And if you don't think they can survive or if you're doubtful about it, you're better to take 50% of your money than end up with zero. Now, Nvidia down 19% is probably an opportunity. Again, not telling you to buy it. Oklo down 61%, years away from revenue. Well, that's a lot riskier, right? You have to have a really serious risk tolerance for that.
So what do you want to avoid? Don't catch falling knives. A stock that falls 50%, it can drop another 50%, and then another 50%, right? So this is just not something that's a smart thing to do. Falling knives are very painful. Just think about it. You drop it, it just hits your foot, right? Painful. So wait for signs of stabilisation. Wait for a base to form. There's a pattern to the stock market. This is a bad one.
And don't average down on losing stocks. Now, if you are buying an index fund, you can average down as much as you want. It works. It does not work for stocks. The risk just gets greater. So if your stock's down 40% and you're going to buy more to lower my average cost, well, that only works if the company's fundamentals are still strong. You're averaging down on a company that's broken. You're just throwing good money after bad because you're not wanting to realise that loss.
And number 3 is revenge trading. Emotions are running pretty high with a lot of people right now. So if you lost money on IonQ, now you want to make a quick buck, right? So you're going to buy the next hot AI stock. And that's emotion. It's not strategy. That's how you turn a 40% loss into a 70% loss.
And then the worst thing I really see with people, I see a lot of people's portfolios, is the lack of risk management. If your stock is 50% of your portfolio right now because everything else went down, that is too much risk in one position. You're going to want to rebalance in my humble opinion. You're going to want to take some of that off the table. I know it's hard to sell at a loss. It's never fun. I've been there. I do it all the time. I do it every week, but I automate it so I don't have to press the button. Much, much easier that way because protecting your capital means you can fight another day. That is always more important than hoping for a miracle.
Now, what are the opportunities? Well, there is always opportunity. The obvious one is high quality AI stocks. Nvidia is 90% cheaper. It's still the dominant AI chip player. Massive revenue, massive profits, massive margins, strong partnerships, right? Proven ability to execute. So it could be interesting.
And then you got the picks and shovels. So during the gold rush, people who made the most money weren't the gold miners. They were the people selling picks and shovels. In the AI boom, who's selling the picks and shovels? It's the power and the utility companies. AI data centres need massive amounts of electricity. So companies that provide reliable power could be good. We have cooling infrastructure stocks. We have data centre REITs. Yes, real estate investment trusts that own data centres, physical locations. You got the semiconductor equipment guys, the companies that make the machines to make the chips. They don't care who wins. They want to make money anyway.
We've also seen very strong defence plays. The government is going to spend more money on wars, right? Looks like the US is going to start another one with Venezuela. Why? Because, well, it's profitable, isn't it? No, no. I'm sorry. I was of course being a cynic. It is for the greater good of humanity somehow. You believe that?
Now, so Santa also says healthcare and biotech are sectors to be in. And why is that? And I've been very bullish on biotech for most of the year. It is actually AI. AI makes biotech much faster, much more profitable, much, much easier.
And then you can always, always go into good quality stocks, value stocks, right? Banks, utilities, consumer staples. I bought a stock 2 weeks ago called Mettler-Toledo. Hasn't done anything since, but it's a high quality company. What do they do? They do precision weighing instruments for the pharma and biotech industry. And then there was a recurring revenue stream from maintaining that equipment. So you want to take a screenshot of these? I'm going to run through these.
Quite quickly, AI stocks here, defensive stocks, a couple of bit more detail on these here. Now, when do we buy? You never know that it's the bottom of the market. Nobody does. So you don't want to be buying because you think it's the bottom of the market. The way I look at it is, a stock does this, it crashes, it goes sideways for a little while. When it starts to recover, that is when I'm looking at it. I don't aim for the bottom. I aim for the recovery, not for the bottom.
Base formation basically. We've done tons of trainings on this in the past. If you haven't joined one of those yet, at least join me on the 27th and I'll walk you through something even more fundamental. That's really incredibly important. And so important, I mean you're putting $1 million in it.
So what else? Join us on Saturday. No, when is the 27th? I never know what day of the week it is. It is a Saturday. Okay, brilliant. Just after the Christmas feast, you can roll yourself off the feast in front of a screen and join me and I will literally walk you through my $1 million experiment.
I'm going to be running that live for the foreseeable future for my students. Every single trade and so on will be shared. The live session here is an explanation of how that works, of what it's all about, why we do it, why it's based, how it's based on the way institutions trade, and a very cool tool that we've built that I think makes that accessible to more people, provided you understand risk management, which is ultimately what this is all about. It's not a promise that you'll make lots of money. It's a promise to be transparent with you, to share this experiment with you, and to teach you structure.
So the link to this is down below in the description, felix/training. And I wish you a beautiful, calm, and relaxed start to the Christmas season. Happy Christmas. If you own Nvidia stock or you've been thinking about buying Amazon, or maybe you just are in an index fund, what I'm about to show you could change everything about your investment strategy. A massive $10 billion.