Advanced Micro Devices Inc. more than doubled its long-term forecast for the server processor market, citing explosive demand from new artificial intelligence workloads that require vastly more computing power.
Advanced Micro Devices Inc. more than doubled its long-term forecast for the server processor market, citing explosive demand from new artificial intelligence workloads that require vastly more computing power.

Advanced Micro Devices Inc. has dramatically reset expectations for the server processor market, with Chief Executive Officer Lisa Su now forecasting a total addressable market of over $120 billion by 2030. The revised outlook, up from a November estimate of $60 billion, is driven by a structural shift in data center architecture for artificial intelligence that could see CPUs become as critical as graphics processors.
“As inference and agentic AI workloads scale, the required CPU compute is far greater than previously anticipated,” Su said on the company’s earnings call. “These tasks, beyond their reliance on GPUs and accelerators, require substantial CPU resources for orchestration, data movement, and parallel execution.”
The bullish forecast came as AMD reported first-quarter revenue of $10.25 billion, a 38 percent year-over-year increase that beat analyst expectations. Growth was powered by its data center segment, which saw sales surge 57 percent to $5.8 billion. The company guided for second-quarter revenue of approximately $11.2 billion, also ahead of consensus estimates.
AMD’s projection signals a new phase in the AI infrastructure buildout, suggesting the computational needs for AI are broadening beyond just training models with GPUs. The forecast directly challenges the dominance of rival Nvidia Corp. and implies a massive, sustained demand cycle that will benefit the entire semiconductor supply chain, from foundry partner TSMC to memory chip manufacturers.
The core driver behind AMD’s revised forecast is the rise of so-called agentic AI — autonomous or semi-autonomous AI bots that perform complex tasks. While AI model training is heavily reliant on GPUs, these new agentic workloads require a significant number of CPUs to manage and orchestrate tasks, leading to a fundamental change in data center design.
According to Su, the industry is seeing a rapid shift in the ratio of CPUs to GPUs. Where data centers were previously built with a ratio of one CPU for every four or eight GPUs, new deployments are moving closer to a one-to-one ratio. In some high-density agent scenarios, Su noted, the number of CPUs could even exceed the number of GPUs. This observation was echoed recently by Intel Corp. CEO Pat Gelsinger, indicating a broad industry trend.
Reflecting this demand, AMD expects its server CPU revenue to accelerate, forecasting year-over-year growth of more than 70 percent in the second quarter.
Alongside the CPU demand surge, AMD announced a major expansion of its strategic partnership with Meta Platforms Inc. The social media giant will deploy up to six gigawatts of AMD’s Instinct GPUs, including a new custom-designed GPU based on the upcoming MI450 architecture. Shipments of the custom chip, co-designed for Meta’s next-generation AI workloads, are expected to begin in the second half of 2026.
This deal represents a significant validation of AMD’s strategy to compete directly with Nvidia in the AI accelerator market. The company is also preparing to launch its next-generation EPYC server processors, codenamed “Venice,” which will use a 2-nanometer manufacturing process and are slated for release later this year.
Despite the torrid growth in its data center business, AMD faces headwinds in its consumer-facing segments. Chief Financial Officer Jean Hu warned that gaming revenue is expected to decline by more than 20 percent in the second half of the year, citing the impact of rising memory and component costs on the PC and gaming markets.
However, investors focused on the larger AI opportunity, pushing AMD shares up more than 6 percent in after-hours trading following the announcements. The results show that while the consumer electronics market faces near-term pressures, the long-term investment cycle in AI infrastructure is proving to be larger and more sustained than even bullish analysts had previously forecast.
This article is for informational purposes only and does not constitute investment advice.