Coatue Management's Philippe Laffont Takes Profits by Selling Stock of AI Giants, Pivots to Taiwan Semiconductor

Philippe Laffont’s investment moves are speaking volumes about where he sees opportunity in the evolving AI landscape. According to Form 13F filings submitted on February 17, 2026, the billionaire manager of Coatue Management—which oversees $40 billion in assets—has been actively rebalancing his portfolio by selling stock of his former top holdings while dramatically repositioning into Taiwan Semiconductor Manufacturing.

The strategic shift represents more than a simple trade. It reflects a calculated decision to take profits from positions that have already delivered exceptional returns while positioning the fund for what Laffont evidently sees as the next critical infrastructure play in artificial intelligence.

Selling Stock to Lock In Gains: The Nvidia and Meta Story

During the fourth quarter alone, Coatue reduced its Nvidia position by 667,405 shares and trimmed its Meta stake by 253,768 shares. Yet these recent moves are merely snapshots of a much larger three-year trend. Since March 30, 2023, Laffont has slashed his Meta holdings by more than half—a cumulative reduction of 4,279,854 shares representing a 53% decrease in that position. His Nvidia exposure is even more dramatic, with an 82% reduction totaling 40,598,682 split-adjusted shares over the same period.

The mathematical reality explains much of this activity. Nvidia’s stock price has surged approximately 1,200% since 2023, while Meta shares have climbed roughly 445%. When positions appreciate at such extraordinary rates, selling stock to harvest gains becomes a prudent portfolio management strategy—especially for a manager overseeing billions of dollars.

Yet there’s another layer to consider. In continuously selling stock of these AI infrastructure leaders, Laffont may be signaling concerns about stretched valuations and the historical pattern of technology bubbles. Every major technological breakthrough from the past three decades has experienced early-stage pullbacks as investor enthusiasm outpaced realistic adoption timelines. While AI infrastructure demand remains robust, significant optimization work likely lies ahead before these technologies translate into sustainable earnings growth.

Taiwan Semiconductor Emerges as the New Portfolio Crown Jewel

The real story in Coatue’s latest 13F becomes apparent when examining what Laffont purchased instead. By adding 556,988 shares during the fourth quarter, Taiwan Semiconductor Manufacturing (TSMC) has now become the fund’s No. 1 position and, in Laffont’s estimation, the leading AI stock worth owning.

TSMC’s ascendancy makes strategic sense. The world-leading chip fabricator has experienced extraordinary demand from the expansion of high-bandwidth memory and GPU production. The company is racing to expand its manufacturing capacity at an unprecedented pace, capitalizing on a structural supply-demand imbalance that should persist as long as GPU orders exceed available supply. This favorable environment supports both TSMC’s order backlog and its pricing power.

What particularly distinguishes TSMC from pure-play AI beneficiaries is its diversified revenue foundation. Beyond its thriving AI semiconductor business, the company manufactures wireless chips for next-generation smartphones, produces advanced semiconductors for Internet of Things applications, and supplies critical processors for automotive systems. While these segments grow more slowly than AI-related operations, they nonetheless provide stable cash flows and a dependable earnings floor—precisely the kind of quality foundation that appeals to sophisticated institutional investors.

The Valuation Argument: Why Laffont’s Pivot Makes Financial Sense

The valuation picture helps explain Laffont’s investment thesis. TSMC trades at a forward price-to-earnings multiple of 21—a reasonable valuation level if the company executes on its projected 31% sales growth in 2026 and 24% growth in 2027. This framework suggests TSMC offers attractive risk-reward dynamics compared to many of its higher-flying AI-adjacent peers.

For a portfolio manager consistently hunting for positions where quality, growth trajectory, and valuation align favorably, TSMC represents precisely that intersection. The company combines the secular AI demand tailwinds that justified years of selling stock in Nvidia, with more balanced valuation metrics and revenue diversification that reduce concentration risk.

The broader lesson from Laffont’s portfolio repositioning extends beyond one fund’s tactical decisions. It illustrates how even billionaire investors with excellent long-term records recognize when to rotate capital—knowing when selling stock from appreciated positions serves the ultimate goal of building lasting wealth across market cycles.