What's happening
At TSMC's 2026 Annual Shareholders' Meeting on June 4, 2026, Chairman and CEO C.C. Wei addressed the company's capacity situation directly, telling shareholders that demand for advanced semiconductors driven by artificial intelligence applications continues to outpace the company's ability to supply. 'It will be a long time before we can meet customer demand,' Wei said. The meeting, held at the Sheraton Hsinchu Hotel in Hsinchu, Taiwan, comes as TSMC forecasts a 30% increase in sales for 2026 and continues to execute on a $165 billion investment commitment in new manufacturing facilities in Arizona.
On the subject of pricing, Wei struck a calibrated tone. 'I'd like to do that... we still need to make money,' he said when asked about potential price increases, while drawing a clear distinction between TSMC's approach and the volatile pricing dynamics seen in the memory chip segment. 'We won't raise prices like memory chips — all of a sudden 400% hikes in the short term. Although I do envy their 80% gross margins, we won't do that. That's not realistic,' Wei stated. The remarks indicate that while TSMC sees room to adjust pricing, it intends to do so gradually rather than through abrupt market-rate moves.
Why it matters for markets
TSMC's 30% projected sales increase for 2026 reflects the scale of AI-driven semiconductor demand flowing through the world's dominant contract chipmaker, which serves major customers including Apple, Nvidia, and AMD across process nodes from 3nm to mature technologies. The company's share price has reached T$2,425, compared to a 52-week low of approximately T$946, underscoring the degree to which capital markets have re-rated the foundry's strategic position over the past year. With a current market capitalization of $2.18 trillion and a P/E ratio of 36.2, TSMC's valuation reflects expectations of sustained earnings growth — a trajectory supported by the projected continuation of roughly 30% profit-sharing growth in 2026, following comparable increases from 2023 to 2024 and from 2024 to 2025.
Wei's pricing commentary carries particular weight for the broader semiconductor supply chain. By explicitly acknowledging a desire to raise prices while committing to avoid abrupt increases, TSMC is signaling that margin expansion, if it materializes, will be incremental rather than disruptive to customer cost structures. For fabless chip designers that depend on TSMC for high-volume advanced-node production, the distinction matters: a gradual pricing environment is more manageable to model into product roadmaps than sudden cost shocks. At the same time, Wei's candid acknowledgment of capacity constraints — and the long timeline to resolve them — confirms that supply tightness will remain a structural feature of the advanced foundry market for the foreseeable future.
The $165 billion Arizona investment represents TSMC's most significant geographic diversification of manufacturing capacity to date. While that buildout addresses long-term supply concerns and geopolitical risk considerations, it does not resolve near-term constraints. The gap between current demand and available capacity, as described by Wei, means that allocation decisions at TSMC continue to carry direct consequences for the product timelines and competitive positioning of its largest customers.
Sectors and assets to watch
The most directly affected tickers are TSMC (TSM), Nvidia (NVDA), and AMD (AMD). TSMC sits at the center of the capacity constraint story as the sole manufacturer capable of producing the most advanced logic chips at scale. Nvidia, with a market capitalization of $4.96 trillion and revenue of $253.49 billion, is among TSMC's largest customers for advanced-node production, with its data center GPU lineup — including the A100 and H100 — manufactured through TSMC's high-volume foundry processes. AMD, with a market capitalization of $796.47 billion and revenue of $37.45 billion, similarly relies on TSMC for its EPYC server processors and Instinct AI accelerators, making foundry capacity allocation a direct variable in AMD's ability to fulfill data center demand.
Beyond these three primary tickers, the supply chain implications extend to the broader ecosystem of fabless semiconductor companies and systems integrators that compete for TSMC's constrained advanced-node capacity. Wei's comments about pricing also have relevance for any company building AI infrastructure, as foundry cost structures feed directly into the bill of materials for AI accelerators and the economics of large-scale data center deployment.
What to watch next
Key developments to monitor include TSMC's quarterly earnings disclosures for confirmation of the 30% sales growth forecast, any formal announcements regarding pricing adjustments for advanced-node capacity, and progress milestones on the Arizona manufacturing buildout. Investors and analysts will also be watching whether Nvidia's and AMD's product roadmaps and delivery timelines reflect any changes in foundry allocation as TSMC manages demand against constrained supply. Wei's explicit framing — that meeting customer demand remains a long-term challenge — suggests that capacity announcements, technology node transitions, and any signals of demand moderation from hyperscale AI infrastructure buyers will be closely scrutinized as leading indicators of how the supply-demand imbalance evolves through the remainder of 2026.