The $16 Billion IT Investing Gamble: Why Warren Buffett's TSMC Bet Became His Most Expensive Departure

When legendary investor Warren Buffett stepped down as Berkshire Hathaway’s CEO in January 2025, he left behind a portfolio that had generated near-incomprehensible returns. His Class A shares achieved a cumulative gain exceeding 6,100,000% under his leadership—a track record built on unshakeable investment principles. Yet even titans of finance aren’t immune to costly mistakes, particularly in fast-moving sectors like IT investing. One pivotal decision to exit Taiwan Semiconductor Manufacturing (TSMC) has revealed just how expensive departing from his core philosophy can be.

The Foundation of Buffett’s IT Investing Philosophy

Before examining what went wrong, it’s essential to understand what made Buffett’s approach to IT investing and broader markets so consistently effective across six decades. His success wasn’t rooted in chasing trends or timing markets—it stemmed from disciplined, time-tested principles that transcended economic cycles.

Buffett’s most defining trait was his commitment to long-term ownership. Rather than viewing stock purchases as short-term trading vehicles, he acquired stakes with the explicit intention of holding them for years, sometimes decades. He recognized that markets experience inevitable boom-and-bust cycles, but grasped a crucial insight: expansion periods dramatically outweigh downturns over extended timeframes. This understanding became especially relevant for IT investing, where technological disruption could create durable competitive advantages.

Value was another obsession. Buffett distinguished sharply between purchasing excellent businesses at fair prices versus mediocre companies at steep discounts. When market sentiment turned pessimistic and prices dislocated from fundamental value—particularly in cyclical sectors—he would patiently accumulate positions.

The third pillar was his focus on competitive moats and industry dominance. In IT investing specifically, Buffett sought companies with sustainable advantages: network effects, switching costs, or unique technological capabilities that competitors couldn’t easily replicate. He preferred owning leaders with pricing power and protected market positions.

Finally, trust mattered immensely. Buffett gravitated toward well-managed businesses that cultivated deep customer loyalty. He understood that corporate trust could evaporate overnight but took years to rebuild—a principle increasingly critical in IT sectors dependent on customer data and security.

The Taiwan Semiconductor Opportunity and Strategic Miscalculation

During the third quarter of 2022, Berkshire Hathaway initiated what appeared to be a textbook Buffett move. The market was experiencing a significant downturn—ideal conditions for identifying price dislocations. The company purchased over 60 million shares of Taiwan Semiconductor Manufacturing, committing approximately $4.12 billion. The thesis seemed sound: TSMC was uniquely positioned as the world’s dominant chip foundry, manufacturing the advanced semiconductors powering artificial intelligence infrastructure.

TSMC’s technological edge was formidable. The company’s specialized chip-on-wafer-on-substrate (CoWoS) process stacked graphics processing units with high-bandwidth memory—precisely what AI data centers required. Apple, Nvidia, Broadcom, Intel, and Advanced Micro Devices all depended on TSMC’s production capacity. The semiconductor was on the cusp of the AI revolution, and TSMC held the critical gatekeeper position.

Yet Buffett’s position in TSMC proved remarkably short-lived—lasting approximately five to nine months. SEC Form 13F filings reveal that Berkshire sold 86% of its stake in the fourth quarter of 2022, completely exiting by the first quarter of 2023. In May 2023 conversations with Wall Street analysts, Buffett articulated his reasoning: “I don’t like its location, and I’ve reevaluated that.” His comments almost certainly referred to the CHIPS and Science Act, enacted in 2022 under President Biden, which incentivized domestic U.S. semiconductor production. Following this legislation, U.S. export controls tightened significantly, restricting advanced AI chip shipments to China.

Buffett apparently feared that similar export restrictions or geopolitical complications would undermine TSMC’s competitive position or growth trajectory. It was a logical concern—but the timing of his exit proved catastrophic.

When IT Investing Trends Overwhelmed Macro Concerns

What Buffett couldn’t have fully anticipated was the explosive acceleration of AI adoption and the resulting semiconductor supply crisis. Nvidia’s GPUs became indispensable for AI model training, creating insatiable demand. Chip backlogs stretched for months, and TSMC found itself racing to expand CoWoS production capacity aggressively month after month.

TSMC’s growth rate didn’t merely accelerate—it exploded. The company’s stock price surged accordingly. By July 2025, TSMC joined the exclusive trillion-dollar market capitalization club, becoming one of only twelve public companies to achieve that milestone.

The numbers tell a stark story. Had Berkshire maintained its initial stake without selling a single share, that $4.12 billion investment would have appreciated to nearly $20 billion as of late January 2026. Instead, Buffett’s premature exit cost his company approximately $16 billion in unrealized gains. That represents a painful reminder of how departing from core IT investing principles—particularly the commitment to long-term holding—can prove extraordinarily expensive.

The Broader Lesson for Long-Term IT Investing

This episode stands out because it violates nearly everything Buffett taught about durable investment success. He purchased during a genuine market downturn when prices reflected value dislocations. TSMC possessed the technological moat he consistently sought. The business had a clear competitive advantage in an emerging, high-growth sector.

Yet short-term macro concerns about geopolitical risk overrode his long-term conviction. Buffett allowed himself to be influenced by policy uncertainty rather than maintaining faith in TSMC’s fundamental competitive positioning within the AI infrastructure ecosystem. It’s a rare lapse in discipline—but an instructive one for anyone pursuing IT investing strategies.

Berkshire’s new CEO, Greg Abel, will likely ensure the company returns to strictly adhering to these investment principles. The costly TSMC experience underscores why long-term, conviction-based IT investing remains superior to reactive trading driven by short-term geopolitical or policy shifts. Markets reward patience and fundamental analysis far more consistently than they reward fear-based tactical exits, particularly in transformative sectors like artificial intelligence where supply constraints and structural demand growth support extended appreciation cycles.