Billionaire Philippe Laffont of Coatue Management has sold Nvidia stock in 10 of the last 11 quarters, unloading over 40 million split-adjusted shares since early 2023. While some profit-taking is expected after a 1,200%+ run, the consistent, massive scale of selling signals deep concern about three converging threats: an AI infrastructure bubble, a erosion of pricing power from custom chips, and an unpredictable China/trade dynamic that menaces a stock priced for perfection.
The conventional narrative on Nvidia (NASDAQ: NVDA) is one of unassailable dominance. Its GPUs are the indispensable pick-and-shovels play of the artificial intelligence (AI) revolution, generating astronomical growth and margins. Yet, a quiet, persistent exodus from one of Wall Street’s most astute investors tells a more complex story.
Philippe Laffont, founder of the $30 billion hedge fund Coatue Management, has been a net seller of Nvidia in 10 of the last 11 quarters, from Q1 2023 through Q4 2025. His stake has contracted from a peak of nearly 50 million shares to just over 9.2 million shares—a disposal of over 40 million split-adjusted shares as the stock vaulted higher. This isn’t a one-time trim; it’s a strategic retreat executed with discipline across years of relentless rally. The market dismisses it as routine profit-taking. A deeper look suggests Laffont is hedging against a future Wall Street isn’t pricing in.
The Profit-Taking Obviousness
Let’s acknowledge the baseline. Coatue first built its mega-stake in Nvidia during the depths of the 2022 bear market. From that initial entry, the stock is up well over 1,200%. For a quantitatively-driven fund like Coatue, locking ingenerational wealth is a rational, textbook move. Laffont has a documented history of taking profits on major winners. On this level, the selling is unremarkable.
But the *scale* and *consistency* of the selling, especially as Nvidia’s market capitalization ballooned past $3 trillion, is unusual. It points to a re-rating of the company’s risk/reward profile that goes beyond simple portfolio management. To understand what might be worrying a mind like Laffont’s, we must analyze the specific vulnerabilities he is likely modeling.
The Three-Pronged Risk Laffont Sees
Our analysis of the selling pattern and the strategic environment points to three interconnected headwinds that threaten Nvidia’s pristine growth narrative.
1. The AI Bubble’s Inevitable Inflection
The first risk is cyclical and historical. History hasn’t been kind to game-changing technologies in their early, hype-driven phases. The dot-com bubble is the canonical example, but the pattern repeats: railroads, radio, the internet itself. The pattern is always the same: explosive investment in infrastructure precedes the period of actual, profitable optimization by years.
This is the AI moment right now. Enterprises and cloud providers are engaged in a GPU arms race, buying Nvidia’s latest Blackwell chips with little regard for near-term ROI. That’s adoption. What’s missing is *optimization*—the phase where businesses figure out how to monetize these massive compute investments efficiently. That lag is the breeding ground for a bubble. When capital expenditure growth inevitably slows from its current torrid pace, Nvidia’s growth will decelerate, and its sky-high price-to-sales ratio, which intimates it’s priced for perfection, will face a harsh revaluation.
2. The Silent War: Custom Silicon Eats Into the Moat
Competition is the most immediate and tangible threat, but it’s not from AMD or Intel in the traditional sense. The danger is from Nvidia’s own customers. Virtually every major tech giant—from Amazon and Microsoft to Google and Meta—has an internal silicon project (AWS Inferentia/Trainium, Google TPUs, etc.).
While these in-house chips are generally less capable than Nvidia’s cutting-edge GPUs, they are dramatically cheaper and, crucially, available without the supply constraints and premium pricing Nvidia commands. This creates a powerful two-tier market: for cutting-edge training, Nvidia has a monopoly. For high-volume, cost-sensitive inference (the eventual bulk of AI workloads), customers have a viable, cheaper alternative. This dynamic is a direct, long-term pressure on gross margin and pricing power, the twin engines of Nvidia’s financial machine. As one analysis noted, Nvidia’s biggest competitive risk isn’t a faster rival chip—it’s these internally developed alternatives.
3. The China Wildcard and Trade Whiplash
The third risk is geopolitical and binary. Nvidia’s revenue from China has effectively been reduced to near zero due to U.S. export controls. But the threat isn’t static. The tariff and trade landscape is highly volatile, and China’s refusal to purchase even the downgraded H20 chip demonstrates a willingness to absorb short-term pain for long-term technological independence. For a company trading at a valuation that assumes flawless, uninterrupted global scale, this is a material risk factor. Any escalation in tech tensions or a successful Chinese domestic alternative would remove a multi-billion-dollar growth vector overnight.
The Valuation Cliff Edge
Individually, each of these risks is debated. Combined, they form a powerful argument for de-risking. Laffont’s selling suggests he believes the market is over-weighting the “AI super-cycle” narrative and under-weighting the timeline and impact of these headwinds.
Nvidia’s valuation leaves virtually no room for error. Its price-to-sales ratio is historically unprecedented for a company of its scale. This pricing implies not just continued dominance, but an acceleration of dominance—a perfect execution path with no competitive, cyclical, or geopolitical setbacks. Such “priced for perfection” setups have a history of sharp corrections when reality, however slightly, disappoints.
The Investor’s Takeaway: Patience as a Strategy
This is not a call to short Nvidia. The company’s technological lead is real and formidable. However, Laffont’s actions serve as a critical case study in investor psychology: the most profitable bets often require holding through extreme volatility, but the most prudent capital preservation may require selling into parabolic moves when fundamental risk/reward deteriorates.
For investors considering a position now, the question is not “Will AI win?” but “Is Nvidia’s stock price assuming AI wins perfectly, on schedule, with no new competition, forever?” The evidence from one of the market’s sharpest minds suggests that assumption is too aggressive. The prudent path may be to wait for a pullback that reflects a more balanced risk assessment, rather than chasing a valuation that already anticipates a flawless decade.
To stay ahead of the narratives that move markets, follow onlytrustedinfo.com for direct, unflinching analysis of the financial trends reshaping your portfolio. We translate breaking developments into the actionable intelligence you need, without the hype or the hand-waving. Read more of our definitive finance coverage here.
