TSMC CEO Warns AI Chip Shortage Will Persist for Years

Industrywide capacity crisis deepens

The world's largest contract chipmaker cannot keep pace with surging demand for AI processors, and the bottleneck extends throughout the entire semiconductor supply chain, TSMC chief executive C.C. Wei told shareholders Thursday.

"It will be a long time before we can meet customer demand," Wei said at the company's annual meeting in Hsinchu, Taiwan. Speaking with reporters afterward, he emphasized that customer appetite for AI chips shows no signs of cooling while his company struggles to fulfill orders. "Customer demand is so high, and we can only support so much," Wei said, adding that TSMC is working to avoid becoming a bottleneck itself.

The capacity crunch reaches well beyond TSMC's own fabrication facilities. Wei noted that suppliers and upstream vendors across the chip industry face similar constraints from AI-related orders, suggesting the shortage reflects systemic limitations rather than isolated production issues.

Pricing pressure and profit margins

Wei acknowledged he would like to raise prices but drew a sharp line between TSMC's strategy and the aggressive rate increases that have disrupted the memory chip market. When asked directly about potential price hikes, he said he would "like to do that" before stressing TSMC has no plans to replicate those tactics. "I envy their 80% gross margins, but I would never do that," he told reporters.

The company maintains its target of revenue growth exceeding 30% for the full year, Wei confirmed.

U.S. expansion faces obstacles

TSMC's $165 billion investment program to expand production capacity in Arizona is progressing, with Wei saying the two sites the company controls should meet its land requirements through the mid-2030s. However, the company's goal of concentrating at least 30% of leading-edge capacity below 2 nanometers on American soil is becoming harder to achieve.

Wei cited slow environmental permitting processes and an insufficient pool of skilled construction labor as key obstacles. Fully satisfying American customers' needs with U.S.-based production will take a "very long time," he added.

Why it matters

TSMC's acknowledgment of prolonged supply constraints signals that AI infrastructure buildouts may face extended delays regardless of capital availability. The industrywide nature of the shortage means hyperscalers and chip designers cannot simply shift orders to alternative suppliers, potentially slowing the pace of AI model deployment and data center expansion across the technology sector.

Workforce investments continue

Employee bonuses tied to profit sharing have increased approximately 30% in each of the past two years, Wei said, with a third consecutive year of similar gains anticipated for 2026.

The company has also been deploying Nvidia AI tools across its fabrication facilities to improve chip yields, reduce defects, and accelerate production timelines in an effort to strengthen manufacturing capabilities.

TSMC stock declined more than 1% on Thursday following the shareholders' meeting.

These details were first reported by AI Watch, citing Reuters and Bloomberg.