Every investor who has been in the markets long enough knows one thing: there are always people, prophets of doom, who will argue that the financial system is broken and a crisis is just around the corner. Roubini, Burry, and Dent are just some of the names who are constantly warning of a huge collapse in the markets. Eventually, they are right, but most of the time, they are wrong.
What I’ve learned as an investor is that, in the long run, you should ignore their comments. Markets are built to rise, and as such, you should use them as an engine to build growth. But lately, the bears’ predictions feel a bit more real, as if a crash is about to happen.
So, are we heading for a stock market crash? Well, I don’t know, and I don’t pretend to know. But I want to check what their claims are all about. For a wider look at who else is sounding the alarm, see our piece on market crash warnings.
In this article
What Burry and the Other Bears Are Actually Saying
Let’s start with Burry. Unlike people who simply keep predicting a market crash, he has put his money behind a bearish view. In an October 11 post, Burry said he had consolidated most of his bearish exposure into put options on the Nasdaq-100 and the VanEck Semiconductor ETF (SMH). He also suggested that the market could be approaching a significant top within the next nine months.
But the interesting part is not just that Burry is betting against the market. It’s why. In his recent writing, he has questioned whether the enormous sums being spent on AI infrastructure will ultimately generate sufficient returns. Basically, he’s saying that major technology companies are investing heavily in data centers, chips, and computing capacity, while other companies are borrowing to participate in the same race. Simply put, Burry’s concern is that the industry may be building much more capacity than it can eventually monetize, leaving investors with expensive assets that fail to deliver the expected returns.
And Burry is not just betting against AI. In recent posts, he also highlighted the growing role of private credit, private equity, and insurance companies in financing data centers. He argued that the growing use of leverage could become a problem if financing timelines don’t match the time these investments need to generate returns. There’s another issue, too: long-term interest rates are rising before these investments have had time to prove themselves.
In other words, Burry’s concern is not simply that AI stocks have gone too far. It’s that the financial structure supporting the boom could become increasingly fragile if spending fails to deliver the expected returns or borrowing costs remain high. Difficult to argue against that.
And Burry is not the only well-known investor questioning the AI boom. Paul Tudor Jones is also warning that 2027 could be a difficult year for stocks, although his argument is somewhat different. He points to a potential surge in equity issuance, including major technology IPOs, alongside concerns about the U.S. fiscal deficit. His argument is that the supply of shares could increase just as the market loses some of the support it has received from corporate buybacks and mergers.
Ray Dalio, on the other hand, is focused more directly on debt and borrowing costs. He has warned that governments and AI companies could end up competing for a limited pool of capital, pushing yields higher and putting pressure on financial asset prices. He estimates that annual net debt supply could increase from approximately $3.5 trillion in 2025 to $4.4 trillion in 2027.
Neither argument is identical to Burry’s, but both point to risks that extend beyond whether AI companies are overvalued.
So, in a nutshell, these guys aren’t just throwing out a warning here and there. They’re seriously betting on a major market correction, or maybe even a crash, and some believe 2027 could be the year it happens. But here’s the thing: they all have positions to defend, and their views reflect their own investment strategies and incentives. That’s worth keeping in mind before jumping on the crash bandwagon.
For Now, Earnings Tell a Different Story
OK, so Burry and his friends are warning that a crisis is coming. But so far, the markets have been remarkably resilient. Yes, we’ve seen pullbacks, including during the US-Iran conflict, and the latest tech sell-off shows that investors are becoming more sensitive to the risks around AI. But we haven’t seen the kind of widespread financial stress that would convince me a major crisis is just around the corner. If I were judging the market by how companies are performing, I’d have to say the bulls still have a pretty good argument.
More importantly, the stock market has kind of lost its connection to economic developments. Over the past few years, it has continued to advance despite elevated interest rates, geopolitical tensions, and concerns about inflation. Sometimes, it almost feels as if inflation and economic growth data are being brushed aside whenever corporate earnings come in strong. And there is a reason for that: ultimately, stocks are claims on future profits, and as long as companies keep growing their earnings, investors have a reason to remain optimistic.
The numbers support that argument, at least for now. According to FactSet’s October 16 earnings update, analysts expect S&P 500 earnings to grow by 29% year over year in the third quarter, marking the third consecutive quarter of growth above 25%. Revenue is projected to increase by 12.6%. It’s still early in earnings season, with only 9% of S&P 500 companies having reported, but 84% of those companies have already beaten earnings-per-share estimates.
Q2 was even stronger, with S&P 500 earnings growing approximately 52% year over year, the highest growth rate since Q2 2021, according to Reuters. However, there’s an important caveat: much of that growth was boosted by gains on AI-related investments at Alphabet and Amazon. Excluding those gains, S&P 500 earnings growth was estimated at 33%. Even with that adjustment, the underlying earnings picture remained strong.
And Burry needs to keep this in mind; AI remains a major part of the story. Technology and communication services are among the sectors expected to deliver some of the strongest earnings growth. Microsoft’s latest reported quarter, for example, saw revenue increase 18% year over year to $90 billion, while diluted earnings per share rose 32%, supported by continued demand for cloud and AI services. And MSFT is not alone; most big and medium-sized tech companies are doing well.
Of course, strong earnings don’t mean every stock is cheap or that the market can’t fall. But they do make it harder to argue that a major crash is inevitable right now. The bears need something to change: earnings expectations need to deteriorate, AI spending needs to stop delivering the expected returns, or financial conditions need to become restrictive enough to put pressure on companies and investors.
Live odds from Polymarket. The first market covers the end of 2026, the second the end of 2027, the year the bears are warning about.
Prediction markets carry a risk of loss, are not available in every country, and are not investment advice. AllinAllSpace may earn a commission if you sign up through our Polymarket link.
So, What Would Actually Trigger a Crash?
Ok, so if the bears are right about 2027, what exactly needs to go wrong for their predictions to become reality?
Surely, something needs to change. As of writing, markets seem far from getting into a crash or even a correction, despite the sharp sell-off in chip stocks this week. Yet, while strong earnings are keeping the bull market alive for now, there are reasons to believe that parts of the AI trade have become vulnerable. The question is whether these risks are enough to trigger a correction or something more serious.
After all, a crash doesn’t happen simply because stocks are expensive. Something needs to break, and in today’s market, those who believe that the next ‘Big Short’ is coming see three potential triggers.
The First Domino: AI Spending Starts to Crack
If there’s one place I’d look for the first cracks, it’s the semiconductor industry. The entire AI boom depends on companies continuing to invest heavily in chips, data centers, and computing infrastructure. But what happens if that spending eventually stops growing fast enough to justify the valuations investors are paying today?
That’s the domino that Burry predicts, and this is not just a hypothetical concern. A recent FactSet analysis found that five major technology companies are expected to see free cash flow close to zero or negative in fiscal 2026 as capital expenditure outpaces their cash flow, raising questions about how much longer they can sustain the current pace of investment. Also, some analysts expect AI infrastructure spending to slow in 2027 as financing constraints become more challenging. In regard to that, the Financial Times reported that analysts at Bernstein consider only about 35% of the planned US data-center pipeline credible, while a tracker counted at least 47 large data-center projects delayed, withdrawn or blocked in 2026.
Reasons for concern? Absolutely. But of course, this doesn’t mean that AI demand is about to collapse. Companies could keep investing, revenues could catch up with spending, and the industry could continue growing. The real concern is that if investment slows while expectations remain high, semiconductor companies could face weaker orders and earnings downgrades. That’s when today’s expensive valuations could become a serious problem.
This is why I’d keep an eye on the Philadelphia Semiconductor Index (SOX). This is the index to watch now and next year, so it’s worth adding to your watchlist.
Recently, it fell nearly 16% between October 6 and October 16, a reminder of how quickly sentiment can change. It bounced back for a few days, but on October 22 it dropped another 6.7% in a single session to close at 10,731.98, its lowest level since July 29 and about 27% below its June 22 record. The sell-off followed fresh worries about AI spending: Lam Research, a chip-equipment maker, warned of supply constraints stretching into 2027, and a filing showed Meta’s future spending commitments had grown to about $825 billion, up from roughly $700 billion in July. That decline does not mean a broader crash has begun. After all, its YTD is still up roughly 50%, so it can be just a correction. What matters is what happens next: if the index continues falling alongside weaker earnings forecasts and signs of slowing AI investment, it could signal that investors are beginning to question the entire growth story.
SOXX is a semiconductor ETF, a fund that holds chip companies and trades like a stock. It is a tradable way to follow the same sector as the SOX index mentioned above, but it is not the same index. Data is delayed. Source: TradingView.
On October 22, our sector rotation tracker showed money moving out of semiconductors and technology and into defensive sectors like consumer staples and utilities. For now, the key question is not whether semiconductor stocks have already started crashing. It’s whether the enormous spending that has supported their rise can continue into 2027, and what happens to the wider market if it cannot.
The Second Domino: Who Is Going to Pay for All This?
This is where Ray Dalio’s argument becomes interesting. In his analysis of the growing supply of debt, he estimates that annual net debt supply could rise from $3.5 trillion in 2025 to $4.4 trillion in 2027, with non-financial corporations alone needing about $1.4 trillion of it. In his view, governments and technology companies are increasingly competing for the same pool of investor capital.
And that could create a problem. If too many borrowers need money at the same time, investors may demand higher yields to lend it. Higher financing costs then eat into cash flow and make new infrastructure projects less attractive. Yes, the Federal Reserve can influence short-term interest rates, but it cannot guarantee that long-term borrowing costs will fall if inflation remains stubborn and the supply of debt keeps growing. I wouldn’t count on the Fed to solve this financing problem, either. By the time it responds to mounting financial stress, some of the damage may already have been done.
We’re already seeing some signs of stress. On October 14, CoreWeave shares fell approximately 15%. The following day, credit-default swap spreads on its debt moved above 1,000 basis points, reflecting serious concern about its credit risk, according to Reuters.
Of course, CoreWeave is just one company, not the entire AI industry. But it illustrates what can happen when a capital-intensive business relies heavily on financing to keep expanding. If the expected returns fail to materialize next year, the consequences can extend beyond shareholders to lenders and other companies in the supply chain. And this, my friends, can indeed create a sell-off.
The Final Domino: Investors Head for the Exit
This is where things could get ugly. If AI spending slows, chipmakers lose orders, and heavily indebted infrastructure providers struggle to refinance, investors may start reassessing the whole sector. And once a sell-off gathers momentum, it doesn’t necessarily stop at the companies that caused it.
And let’s not forget how much optimism is already reflected in the market. Many AI-related stocks have delivered enormous gains, and investors have grown accustomed to strong earnings and ambitious spending plans. If that story starts to unravel, some investors will take profits, others will cut their losses, and leveraged investors may have little choice but to sell.
That doesn’t mean profit-taking alone will cause a crash. But when it coincides with deteriorating earnings expectations, rising financing costs, tighter credit conditions, and, most importantly, a loss of investor confidence, it can amplify the decline. The danger is that what begins as a correction in semiconductor stocks may turn into a broader reassessment of risk across the market.
This is the chain of events that can trigger a crash: AI spending disappoints, financing becomes more expensive, and investors start pulling back. If all three happen together, a sector sell-off could become something much bigger. If earnings remain strong and AI demand continues to grow, however, the bears may have to wait longer than they expect.
That’s the difference between saying the market is vulnerable and claiming that a crash is inevitable. For Burry and his fellow bears to be right about 2027, we need to see these risks move from warnings to actual damage in company earnings, financing conditions, and the wider market.
For that reason, right now, more than ever, earnings reports and bond yields across the globe are the two metrics every investor needs to watch, both now and next year.
Buy, Sell or Hold?
So, what are we saying here? Should you sell your portfolio or part of it? Not necessarily.
The broader market, as of writing, is holding up despite the chip sell-off, and we are not seeing any signs of an imminent crash. In fact, since the COVID-19 correction (please don’t tell me it was a crash), we have constantly heard that the markets are overvalued. But the reality has shown something different. As long as the market doesn’t fall, there’s no reason to sell.
For now, despite the bears’ predictions, I’d hold my portfolio, but with a finger on the button. You might want to reduce exposure to AI and semiconductor stocks and ETFs, but the general sentiment remains positive. What I’d do is keep an eye on developments in the AI sector, including earnings reports, valuations, and AI spending. Between now and the end of next year, every minor correction could turn into a major sell-off, so the risk is real and must be closely monitored.
This article is for informational and educational purposes only and does not constitute financial or investment advice. Investing in stocks, ETFs and options involves a high risk of loss, and forecasts may not come true. Sources: FactSet, Reuters, Microsoft, Financial Times, Stocktwits, 24/7 Wall St, Mettis Global, Livewire Markets, TradingKey and Cappnotes. Polymarket odds change constantly. Figures accurate as of October 22, 2026.