AMD plans to substantially increase its supply of AI chips in 2027 as demand for artificial intelligence infrastructure continues to rise, CEO Lisa Su said during a visit to Taiwan on October 6.
Su said AMD has already increased supply during 2026 and expects a much larger expansion next year. She said the company is working with manufacturing and supply-chain partners to increase production of both CPUs and GPUs.
AMD is planning further capacity
The immediate challenge is not simply designing faster chips. AMD also needs enough manufacturing capacity, advanced packaging and high-bandwidth memory to actually ship those chips at scale. That is why Lisa Su’s trip to Taiwan matters. She met with supply-chain partners including Foxconn and is expected to meet TSMC, AMD’s key manufacturing partner.
AMD is also working with memory suppliers as demand for AI infrastructure puts pressure on the wider semiconductor supply chain. South Korea’s Samsung and SK hynix are among the major memory companies the industry depends on.
Su’s comments suggest AMD is treating this as a multi-year capacity problem rather than a short-term production bump. She said the company is planning three to five years ahead because it sees very high demand for advanced computing capacity.
AI demand is reshaping the chip market
AMD’s announcement is part of a much larger shift in the economics of computing. Training and running increasingly capable AI systems requires large amounts of compute, memory and networking, turning semiconductor capacity into a strategic constraint for cloud providers and AI companies.
AMD’s own financial filings describe the AI infrastructure buildout as a major technology investment cycle. The company has been positioning its data-center CPUs, GPUs and software platform as pieces of a broader AI computing stack, rather than relying on a single accelerator product.
That creates an interesting opportunity for AMD, but it also raises the bar. Increasing supply does not automatically mean AMD will capture the demand. The company still has to execute on manufacturing, packaging, memory availability, software support and customer deployments while competing with Nvidia and other suppliers.
Su also used the Taiwan visit to stress that the AI boom comes with responsibilities around safety. That is a useful reminder that the semiconductor race is ultimately enabling a much larger technology transition, and the infrastructure decisions being made now will shape how quickly that transition can scale.
What to watch next
The important question is whether AMD can turn strong AI demand into sustained shipments. Watch its manufacturing discussions with TSMC, memory supply agreements, advanced packaging capacity and the pace at which customers deploy AMD accelerators and server CPUs.
Our view is that the 2027 supply increase is significant because it shows the AI chip race is increasingly becoming a supply-chain race as well. Companies can have strong products and still lose opportunities if wafers, packaging or memory cannot arrive when customers need them. AMD’s ability to expand all of those layers together may be just as important as the performance of its next-generation chips.
Sources and further reading
Reuters: AMD plans to substantially increase supply in 2027, CEO says
