Peter Thiel fully exited his Nvidia position in Q3 2025, selling about 537,742 shares. This stake accounted for almost 40% of his investment fund. Thiel’s macro fund drastically reduced overall U.S. equity exposure, trimming holdings from roughly $212 million to $74 million. His portfolio now concentrates mostly on Tesla, Microsoft, and Apple.
SoftBank, once Nvidia’s largest shareholder, sold all 32.1 million shares, valued at $5.8 billion, in October 2025. This move came as SoftBank redirected capital toward new AI ventures like OpenAI and a planned trillion-dollar AI manufacturing hub.
Earlier, Michael Burry also reduced exposure to Nvidia and AI-related stocks, citing stretched valuations and speculative risks.
The exits by these titan investors prompted steep falls in Nvidia’s share price and the broader tech sector. NVIDIA shares slipped following SoftBank’s sale, and SoftBank’s own stock fell 15% in just three days.
Experts view these moves as a data-backed caution signal about AI valuations. Nvidia’s rapid ascent to a trillion-dollar-plus valuation reflected sky-high expectations fueled by explosive demand for AI chips and data center growth. Still, such valuations appear increasingly disconnected from traditional financial fundamentals.While exiting Nvidia, SoftBank is not abandoning AI but reinvesting into new projects with expansive AI infrastructure potential. Peter Thiel’s portfolio shift to Tesla, Microsoft, and Apple suggests a measured emphasis on more diversified, robust tech companies amid current market volatility.The collective withdrawal by these high-profile investors raises the critical question: Is the AI-driven stock rally approaching a bubble burst? Although fundamentals remain strong for leading AI firms, this concentration of exits is fueling debate about overheated valuations and a possible market cooldown ahead.
For investors and market watchers, this data provides an essential signal to re-assess AI stock risks, valuation levels, and portfolio balance in light of recent market dynamics and insider moves.
Across all three moves, one theme is impossible to ignore: valuation stress. The AI sector has seen historic inflows, historic optimism, and historic price expansions. But demand growth doesn’t always translate into sustained profit growth. Early adopters burn cash before they make it. Data centers cost billions. Electricity demand is exploding. Margins get thinner as competition rises. And if earnings fail to match expectations, the correction could be sharp, fast, and painful.
Why are top investors suddenly pulling money out of AI?
Some of the biggest names in global finance are stepping back from AI — not because they don’t believe in the technology, but because they worry that valuations have run far ahead of reality. These moves aren’t minor hedges. They are full-scale exits and billion-dollar bets against the sector.
Michael Burry, famous for predicting the 2008 crash, has placed sweeping put options against major AI-exposed companies. He is effectively betting that their current financial projections are inflated and that future earnings won’t justify today’s sky-high prices. That is not a small claim — it’s a direct challenge to the foundation of the AI boom.
Peter Thiel, another heavyweight known for early bets on game-changing technology, has liquidated his entire position in a leading AI chip maker. For a long time, this stock made up a massive share of his portfolio. Exiting completely suggests genuine concern about the sustainability of current valuations.
SoftBank has also moved out of major AI chip holdings, choosing instead to redirect capital across a broader AI ecosystem. This isn’t simply profit-taking. It’s a signal that the firm sees structural risk in one corner of the AI market — and opportunity in others.
Together, these moves paint a clear picture: the smartest money in the room thinks the AI story is entering a risky phase.
Michael Burry warning that AI profits are being overstated
One of the biggest concerns raised by Burry is that companies powering the AI boom may be overstating their profitability by using accounting choices that make short-term earnings look better. He believes the useful life of AI hardware — especially costly data-center equipment — is being stretched too far.
When companies extend depreciation schedules, their official earnings jump, even if cash flows don’t actually improve. According to Burry’s logic, this could mean that investors are valuing AI companies on inflated, non-durable profitability.
If he’s right, a large chunk of the sector’s earnings could eventually be revised downward. That would force a painful correction, not because AI demand collapses, but because the math supporting current stock prices becomes impossible to defend.
Burry isn’t saying AI is fake. He’s saying the expectations built around it may be unrealistic. And historically, bubbles don’t burst because technology isn’t real — they burst when prices disconnect from what companies can actually earn.
This is why his warning is being taken seriously. It isn’t about hype. It’s about numbers.
Why did Peter Thiel walk away from one of the strongest AI winners?
Thiel’s exit is especially important because he understands tech cycles better than almost anyone. He has a long history of spotting early trends — and stepping aside when valuations no longer make sense. If he sells completely, it usually means he sees either a peak or a shift.
Thiel has hinted that the market is behaving like the late stages of a classic bubblewhere investors stop evaluating fundamentals and start pouring money into a sector simply because prices keep rising. That kind of momentum cycle can last long — until something breaks.
His concern is simple: AI demand is real, but not every company will turn it into long-term profit. And the few that do may take years to justify their current prices.
His full exit suggests he believes risk is now higher than reward. When someone with Thiel’s track record steps back so aggressively, many investors wonder whether he’s seeing something that others are ignoring.
It doesn’t mean AI is going away. It means the stock market version of AI may be entering a dangerous stage.
Is SoftBank exiting AI or shifting to a new strategy?
SoftBank’s decision to sell its massive AI chip holdings has sparked intense debate, but the move is more nuanced than a simple “exit.” SoftBank isn’t abandoning AI — if anything, it’s doubling down on a broader, more physical AI future: robotics, autonomous systems, and infrastructure.
However, the timing raises important questions. Why sell now? Why take profits at this exact moment? Why redeploy capital instead of holding?
SoftBank appears to believe that while AI will grow, the next wave of winners won’t look exactly like the current leaders. They are placing massive capital into AI platforms, data-center projects, and intelligence infrastructure — areas with potentially longer-term payouts.
This could signal that the firm thinks the chip-driven rally is overheatingand the most explosive gains may already be behind it. They want to be early in the next big AI phase rather than the tail-end of the current one.
In short, they are not abandoning AI. They are repositioning for what they believe comes next.
So — is the AI bubble really about to burst?
Some kind of correction now looks increasingly likely. Too many signals are converging: overstretched valuations, accounting red flags, massive insider exits, and shifting capital strategies. That doesn’t mean AI is a fad — it means the market built around AI may be pricing in more than reality can deliver in the short run.
But a full collapse? That is far less certain. Unlike the dot-com bubble, AI has real adoption, real infrastructure, and real use cases across industries. Companies are spending billions not on ideas, but on hardware, models, and deployment.
The most realistic outcome is not a dramatic crash, but a repricing. A cooling-off. A shakeout where strong companies survive and weaker players get punished.
The bubble may not burst — but the air is definitely starting to leak out.