TSMC CEO: Chip Supply Will Trail AI Demand for Years

Supply constraints to persist despite expansion

Taiwan Semiconductor Manufacturing Co. expects chip supply to lag behind AI-driven demand for years, even as the world's largest contract chipmaker brings new manufacturing capacity online in the United States.

CEO C.C. Wei delivered the assessment at TSMC's annual shareholder meeting in Hsinchu, Taiwan, on Thursday, according to Bloomberg. Wei said the company cannot fulfill demand from American customers despite expanding U.S. production facilities, underscoring the scale of computational requirements driving the current AI boom.

The supply-demand imbalance positions TSMC for continued strong financial performance. Wei reiterated the company's forecast for revenue growth exceeding 30% in 2026, a projection that reflects both robust customer orders and the chipmaker's pricing power in a constrained market.

Why it matters

TSMC's multi-year supply shortfall signals that infrastructure bottlenecks—not just algorithmic innovation—will shape the pace of AI deployment across industries. Companies racing to build AI capabilities face extended lead times for the advanced chips that power large language models and other compute-intensive applications. This constraint affects everything from cloud providers expanding data center capacity to enterprises implementing on-premises AI systems. For technology leaders, TSMC's outlook means strategic chip procurement and vendor relationships become critical competitive factors, not just technical architecture decisions.

Capacity expansion meets surging requirements

TSMC has invested heavily in new fabrication plants, including facilities in Arizona that represent the company's most significant U.S. manufacturing presence. Yet Wei's comments indicate these expansions, while substantial, still fall short of meeting the aggregate demand generated by AI workloads.

The gap between supply and demand reflects the technical complexity and capital intensity of advanced semiconductor manufacturing. Leading-edge chip production requires years of planning, billions in investment, and highly specialized equipment and expertise. TSMC's most advanced nodes—the manufacturing processes that produce the chips powering frontier AI models—cannot be quickly scaled.

For customers ranging from hyperscale cloud providers to AI startups, the prolonged shortage means continued competition for limited manufacturing capacity. It also reinforces TSMC's central position in the global technology supply chain, particularly for companies developing and deploying AI systems at scale.

The details were first reported by Bloomberg.