AMD plans to substantially increase its chip supply in 2027 as the US semiconductor company races to keep pace with surging demand for artificial intelligence hardware, CEO Lisa Su said.
Su made the comments in Taipei on Oct 6 during a visit focused on securing additional production capacity across AMD’s supply chain. The company has already increased output during 2026, but demand continues to exceed available supply, particularly as businesses and data-centre operators expand their AI infrastructure.
“We’ve been able to increase our supply as we’ve gone through 2026, and we’re going to substantially increase our supply in 2027,” Su told reporters.
The comments underline the scale of the AI hardware boom — and the growing pressure on chipmakers and their manufacturing partners to secure enough advanced processors, memory and packaging capacity to satisfy customers.
AMD Races to Keep Up With AI Demand
AMD has emerged as Nvidia’s closest major competitor in the market for AI graphics processors, while also maintaining a strong position in server central processing units, or CPUs.
The company’s market value recently surpassed $1 trillion as investors continued to bet on its expanding AI business.
But the rapid growth has also created a supply challenge.
Su said AMD has been increasing supply throughout 2026, yet demand remains higher than what the company can currently provide. She described demand for computing products as “very, very high” for the next several years and said AMD needs additional advanced wafer capacity.
The situation is particularly important because producing advanced chips requires capacity to be planned years before the processors reach customers.
AMD has therefore expanded its supply-chain planning horizon from roughly one to two years to between three and five years, Su said.
Taiwan at the Centre of AMD’s Expansion
Su’s visit to Taiwan was aimed largely at making sure AMD’s manufacturing partners can scale up alongside the company’s expected demand.
During the trip, she met representatives from Foxconn as well as other technology companies including Acer, Asus and Quanta. She was also scheduled to meet Taiwan Semiconductor Manufacturing Co, or TSMC, in Hsinchu.
TSMC is particularly important to AMD because it manufactures many of the company’s most advanced processors.
Su said AMD needs more advanced wafer capacity but did not directly answer questions about whether TSMC had discussed a possible investment in Texas with the chipmaker.
AMD also plans to increase its investment in Taiwan’s semiconductor supply chain beyond the $10 billion commitment it announced in May, according to Su. She did not provide a new figure.
The planned expansion illustrates how AI demand is affecting the semiconductor industry well beyond the companies designing the processors.
Foundries, advanced-packaging providers, substrate manufacturers and memory-chip producers must all expand in coordination if additional AI processors are to reach customers.
Memory Supply Is Another Challenge
AMD’s supply push is not limited to processors.
The company is also working closely with memory-chip manufacturers to ensure it has enough components for its growing AI business.
Su was scheduled to travel to South Korea after her Taiwan visit to meet with major memory companies Samsung Electronics and SK Hynix. South Korea is home to two of the world’s most important suppliers of memory used in modern computing systems.
High-bandwidth memory, or HBM, is particularly important for AI accelerators because it allows processors to access large amounts of data at extremely high speeds.
The industry’s rapid shift towards AI computing has placed significant pressure on HBM production as chipmakers and data-centre operators build increasingly powerful systems.
AMD’s efforts therefore involve securing capacity across several parts of the semiconductor supply chain rather than simply ordering more wafers.
Server CPU Demand Is Also Growing
Although AI graphics processors attract much of the attention, AMD is also benefiting from rising demand for server CPUs.
The company said earlier this year that its server CPU supply chain had been tight, partly because demand was stronger than expected.
During AMD’s second-quarter earnings call, Su said the company expected the server supply situation in 2027 to be better because demand would be more accurately forecast. She also said AMD expected server-related growth of more than 70% year over year, while leaving open the possibility of stronger growth depending on market conditions.
That creates another major opportunity for AMD as data centres increasingly require both GPUs for AI workloads and CPUs to manage broader computing tasks.
The rise of AI agents could further increase demand for CPUs because such systems need processors to handle data processing, tool use and other workloads surrounding AI models.
AMD’s Supply Chain Is Becoming a Long-Term Bet
The semiconductor industry’s current capacity crunch highlights a fundamental feature of AI infrastructure: increasing supply cannot happen overnight.
Building new semiconductor manufacturing capacity requires significant investment and years of planning.
AMD is therefore working further ahead with suppliers to ensure that wafer fabrication, advanced packaging and other components become available at roughly the same time.
Su said the company is now planning three to five years into the future because of the exceptionally strong demand outlook.
The longer planning cycle could also help AMD avoid some of the supply constraints it experienced in 2026.
During its second-quarter earnings call, the company said its server CPU supply had been tight during the first half of the year because much of the demand had not been forecast far enough in advance.
Competition With Nvidia Remains Intense
AMD’s expansion comes as it tries to gain ground on Nvidia, which remains the dominant supplier of AI accelerators.
The competition is increasingly extending beyond individual chips.
Technology companies and cloud providers are looking for complete AI infrastructure that combines GPUs, CPUs, networking, memory and advanced software.
That could create opportunities for AMD because it can supply multiple components of a data-centre system. At the same time, Nvidia’s expanding portfolio means the competition is becoming more intense.
AMD’s ability to increase supply could therefore be just as important as the performance of its individual processors.
If customers cannot obtain enough chips, even strong products may struggle to translate demand into revenue.
AI Boom Driving a Broader Semiconductor Expansion
AMD’s plans are part of a much larger investment cycle across the global technology industry.
Cloud providers and technology companies are spending heavily to build data centres capable of training and running increasingly sophisticated AI models.
That demand is putting pressure on advanced semiconductor manufacturing, high-bandwidth memory and energy infrastructure at the same time.
For AMD, the immediate priority is ensuring that its manufacturing network can expand quickly enough.
“We can definitely use more,” Su said when discussing the company’s ability to increase supply.
The statement captures the central challenge facing the chip industry: demand for AI computing is growing so rapidly that even substantial increases in production may not be enough.
For now, AMD is betting that expanding its supply chain years ahead of time will allow it to capture more of that growth.
And with the company expecting another major increase in chip availability in 2027, the next year could become a crucial test of whether AMD can turn surging AI demand into sustained gains against its much larger rival.
