• Bank of America raises its 2030 server CPU total addressable market forecast to over $170 billion from $125 billion, driven by agentic AI.
  • Analyst Vivek Arya highlights that AI systems planning and executing multi-step tasks will boost CPU usage alongside accelerators.
  • Price targets raised for AMD, ARM, and Intel, signaling a broader semiconductor demand expansion.

AI-Driven Demand Reshapes CPU Outlook

Bank of America has significantly upgraded its long-term view on server CPUs, projecting the total addressable market will exceed $170 billion by 2030, up from a prior estimate of $125 billion. The revision, led by analyst Vivek Arya, hinges on the rise of agentic AI—autonomous systems that can plan and execute complex, multi-step tasks. Unlike earlier AI workloads that relied heavily on GPUs, these new applications require robust CPUs to coordinate and manage workflows, increasing demand for higher-core-count processors.

“Agentic AI is shifting the compute balance,” Arya wrote in a note to clients, arguing that CPUs will play a more central role in data centers. The bank also raised price targets for AMD, ARM Holdings, and Intel, reflecting optimism about their server CPU portfolios. AMD’s EPYC line, Intel’s Xeon processors, and ARM-based designs from partners like Amazon’s Graviton and Google’s Cobalt are expected to benefit as hyperscalers upgrade infrastructure.

Industry Context and Implications

The forecast comes amid a broader AI-driven capex cycle, with hyperscale operators increasing spending on data centers. BofA’s upgrade suggests that the CPU market, long overshadowed by GPU growth in AI discussions, is poised for renewed expansion. The revised TAM also accounts for higher average selling prices as chips become more specialized for AI coordination tasks.

“We see a multiyear growth trajectory for server CPUs,” Arya said, noting that enterprise refresh cycles and ARM adoption will further fuel demand. The note follows similar bullish commentary from other banks, reflecting a consensus that AI workloads will require more balanced compute architectures.

Reached for comment, representatives from AMD, Intel, and ARM declined to discuss the analyst report. The stock moves suggest investors are pricing in the upside, though execution risks remain, particularly for Intel amid its turnaround efforts.

Correction: A previous version of this article misstated the prior TAM figure. It has been updated to $125 billion.