BofA raises its 2030 server CPU market outlook, citing agentic AI demand and widening the growth frame for AMD and ARM

Bank of America raised its targets on AMD and ARM and, in the process, broadened its view of how much demand could ultimately be created for server processors. In a note reported by Yahoo Finance, the bank increased its estimated 2030 total addressable market (TAM) for server CPUs to more than $170 billion, compared with an earlier estimate of $125 billion.

The core of the update is the bank’s argument that “agentic AI” could drive a near-fivefold expansion in CPU demand. Agentic AI generally refers to AI systems that can plan and take actions across tasks, rather than only responding to a single, narrowly defined prompt. In the bank’s framing, that shift increases the amount of compute needed to run and coordinate longer or more autonomous workflows in data centers.

Under that view, server CPU usage becomes a larger piece of the AI infrastructure stack, not just a supporting component. While the update was described as affecting both AMD and ARM, the practical implication is that any platform benefitting from higher CPU content per server, more frequent scaling cycles, or broader adoption of workloads that require ongoing orchestration could see demand expectations move upward.

BofA’s revised TAM estimate is also notable because it places a specific growth timeline on the server CPU category. The bank moved from $125 billion to above $170 billion for 2030, an increase that suggests it expects both more servers and/or higher CPU intensity, as AI workloads spread from experimentation into production and require sustained execution.

For AMD, the change matters because the company sells x86 server processors and also participates in the broader ecosystem that data-center customers use to deploy AI and high-performance computing workloads. In analysts’ conversations, server CPUs are often treated as a volume driver tied to data-center capex cycles, while AI accelerators handle specific tensor and inference workloads. BofA’s emphasis on agentic AI implies that CPUs could play a larger role in overall AI compute demand, at least relative to prior assumptions.

For ARM, the linkage is more indirect but still relevant. ARM’s licensing model and processor designs power a wide range of chips that compete in data centers, including server CPUs based on ARM architectures. By raising overall server CPU TAM and connecting it to agentic AI, BofA is effectively reinforcing a “bigger pie” thesis for CPU supply and platform adoption, rather than treating CPU demand as static.

The bank’s reported note did not provide additional granular disclosures in the Yahoo Finance write-up, such as the specific assumptions behind the “near-fivefold” CPU demand expansion, how much of the TAM increase BofA attributes to market share shifts versus total market growth, or whether the revisions are linked to particular product ramps or customer qualification cycles. It also did not outline new guidance from AMD or any ARM-related announcements at the company level.

Going forward, investors will likely watch for follow-through in two areas. First, any indicates from major data-center operators about workload mix changes toward more autonomous or agentic systems could validate the bank’s demand narrative. Second, AMD-specific updates such as processor roadmap progress, platform design wins, and broader evidence of CPU utilization changes in AI-oriented deployments could determine whether the market expansion thesis translates into near-term results.