ARK Invest Bets on AI Hardware After Meta Earnings Miss Sparks Tech Selloff

Cathie Wood’s ARK Invest aggressively increased its holdings in semiconductor giants following a disappointing earnings report from Meta. The move signals a strategic shift toward the foundational hardware of artificial intelligence during a period of heightened market volatility.

On July 29, Meta reported Q2 2026 earnings per share of $6.18, falling short of the Wall Street consensus estimate of approximately $7.18. The miss triggered a broader selloff across the technology sector, creating a buying opportunity that ARK Invest leveraged between July 28 and July 30. During this window, the firm purchased between 73,000 and 79,000 shares of Nvidia, valued between $14 million and $15.5 million, and invested an estimated $14 million to $20 million into Taiwan Semiconductor Manufacturing Company (TSMC).

The Infrastructure Bet

This rotation comes as the AI market grapples with the massive capital expenditures required by "Hyperscalers." While Meta's earnings miss caused short-term panic, the company simultaneously raised its 2026 capital expenditure guidance to a range of $130 billion to $145 billion. This suggests that despite immediate profit fluctuations, the commitment to building AI infrastructure remains aggressive.

TSMC's own financial performance underscores the strength of this hardware demand. The foundry recently reported quarterly revenue of $40.2 billion, marking a 34% increase year-over-year. For investors like Wood, these figures provide a fundamental justification for prioritizing the companies that design and manufacture the chips over the platforms that deploy them.

Why the 'Picks and Shovels' Strategy Matters

ARK’s trade highlights a classic "picks and shovels" investment approach. By moving into the essential hardware layer, the firm is betting that the AI build-out is still in an early, infrastructure-heavy phase. In this model, the companies providing the necessary tools—Nvidia for chip design and TSMC for fabrication—are viewed as safer, high-growth bets than the software giants spending the capital.

This strategy suggests that even when AI-driven software companies miss expectations, the underlying demand for the hardware required to build those systems remains robust. The risk shifts from whether AI will be profitable for a single platform to whether the entire industry will continue to require massive amounts of compute power.

What to Watch

Market observers will now look to see if other institutional investors follow ARK's lead in rotating away from general big-tech platforms toward semiconductor foundries. The primary tension remains the gap between the soaring capital expenditures of the Hyperscalers and the actual earnings returns generated by AI services. Until that gap closes, the hardware providers remain the primary beneficiaries of the AI arms race.