Philippe Laffont's Latest Portfolio Reshuffle: TSMC Overtakes Nvidia as His New AI Stock Leader

On February 17, 2026, institutional investors managing over $100 million disclosed their quarterly holdings through SEC Form 13F filings. Among these filings, one particularly caught the market’s attention: the investment moves by Philippe Laffont’s Coatue Management, a $40 billion hedge fund known for backing transformative technologies, especially in artificial intelligence.

The story here isn’t just about what Laffont bought—it’s about what his fund strategically sold off. After spending more than a decade loading up on Nvidia and Meta Platforms, the billionaire investor has made a decisive pivot. His latest filing reveals a wholesale repositioning that signals where the smart money believes AI demand will be most concentrated in the coming years.

The Strategic Retreat: Why Laffont Trimmed Nvidia and Meta Holdings

Philippe Laffont’s action was unmistakable. During the fourth quarter alone, Coatue Management reduced its Nvidia position by 667,405 shares and trimmed Meta’s holdings by another 253,768 shares. But the quarterly numbers only tell part of the story.

Since March 30, 2023, when AI became the dominant investment thesis, Laffont’s Meta stake has shrunk by more than half—specifically, by 4.3 million shares, representing a 53% reduction. His Nvidia position has been cut even more aggressively, with an 82% pullback totaling over 40 million split-adjusted shares.

On the surface, this seems counterintuitive. Nvidia’s graphics processing units remain the gold standard for AI compute power, while Meta’s social media properties are virtually unmatched in scale. The stock market has rewarded both handsomely: Nvidia shares have climbed roughly 1,200% since the start of 2023, while Meta has surged approximately 445%.

For most investors, such gains would feel like vindication. For Philippe Laffont, it was a signal to cash in and redeploy capital elsewhere. This suggests one of two things: either the billionaire investor sees warning signs of an AI bubble (historically, transformative technologies experience early-stage corrections), or he’s identified an even more compelling opportunity in the chip supply chain itself.

Taiwan Semiconductor Becomes Coatue’s Flagship AI Bet

The answer emerges in the same 13F filing. Purchasing an additional 556,988 shares in the fourth quarter, Taiwan Semiconductor Manufacturing Company (TSMC) is now the crown jewel of Laffont’s portfolio—the new number one holding.

This isn’t a surprise given TSMC’s position in the AI ecosystem. The Taiwan-based chip manufacturer has been ramping production at breakneck speed, specifically expanding its high-bandwidth memory capacity to meet the voracious demand for GPU-packed processors. As long as demand for advanced chips exceeds supply—a condition that’s likely to persist for years—TSMC maintains substantial pricing power and a healthy backlog of orders.

For Philippe Laffont, the appeal is multi-layered. TSMC isn’t merely riding the AI wave; it’s the foundry underpinning the entire AI infrastructure buildout. Every AI accelerator, every advanced processor, ultimately depends on the manufacturing expertise and capacity that TSMC provides.

Beyond AI: Why TSMC Offers More Than Just an AI Play

What makes TSMC particularly attractive to sophisticated investors like Laffont is its revenue diversification. The company generates substantial sales from wireless communications chips destined for next-generation smartphones, advanced processors for Internet of Things applications, and automotive electronics.

While these segments lack the explosive growth trajectory of AI-related operations, they serve as ballast—providing steady cash generation and a revenue floor that insulates the company from an isolated AI market slowdown. This is precisely the kind of structural stability that appeals to experienced allocators who’ve seen multiple tech cycles.

The valuation picture reinforces this logic. Trading at a forward price-to-earnings multiple of 21x, TSMC appears reasonably priced if the company achieves or surpasses consensus estimates calling for 31% sales growth in 2026 and 24% growth in 2027.

The Larger Investment Thesis

What Philippe Laffont’s moves reveal is a sophisticated understanding of AI economics. Rather than betting on AI’s direct beneficiaries—the consumer platforms and software companies racing to deploy generative AI—he’s positioned Coatue deeper in the infrastructure stack where capital requirements are massive and competitive moats are formidable.

The decision to systematically reduce exposure to Nvidia and Meta, simultaneously with raising TSMC, isn’t a bet against those companies. It’s a bet on where the structural advantages are most durable and where valuations relative to growth prospects offer the most compelling risk-reward profile.

For investors following the moves of elite allocators like Philippe Laffont, the message is clear: when a $40 billion fund reshuffles a decade-old portfolio, it’s worth understanding not just what was bought, but more importantly, why the old thesis was reassessed.