AMD CEO Lisa Su says the company plans to “substantially increase” its supply in 2027 and needs “more advanced wafer capacity”, as she meets Foxconn and TSMC
Context & Ripple Effects
AMD’s planned 2027 supply expansion follows a period in which MI300 supply was expected to stay tight through 2025 and the company reported rapid growth in AI-chip sales. Its dependence on leading-edge manufacturing has been a recurring constraint, including earlier discussion of tight TSMC wafer supply.
TSMC has separately committed to expand Arizona capacity under a $100 billion U.S. investment plan, framing demand as a multiyear trend. Su’s meetings with TSMC and Foxconn put AMD’s supply target in the context of securing production capacity and coordinating the hardware supply chain.
First-order effects
- AMD must secure additional advanced-wafer allocations to deliver the substantially higher 2027 supply it is planning.
- TSMC gains a clearer demand signal from a major AI-chip customer, while Foxconn is drawn into AMD’s planning around scaling hardware output.
Second-order effects
- AMD’s capacity requirements intensify competition for TSMC’s advanced-wafer output, making supply commitments a more important factor in AI-chip availability.
- TSMC’s planned Arizona expansion becomes more strategically relevant to customers such as AMD whose product roadmaps depend on additional leading-edge capacity.
Third-order effects
- If major AI-chip vendors continue to reserve production further ahead, semiconductor competition shifts further from chip design alone toward contracted foundry capacity and manufacturing coordination.
- The AI infrastructure buildout is making wafer availability a product constraint: companies with dependable access to advanced manufacturing can translate demand into shipped systems more reliably.
The trend: AI-chip suppliers are moving toward longer-horizon capacity planning as advanced-wafer availability increasingly sets the ceiling on product supply.