Arm says demand for its new AGI CPU is surging as “agentic AI” takes off, prompting the company to significantly raise its long-term revenue expectations for the chip.
Unveiled in March, Arm’s AGI CPU was built specifically for the growing needs of agentic AI workloads, where systems must coordinate tasks, tools, and models efficiently at scale. The release also signaled a notable shift in Arm’s strategy. Rather than remaining primarily an IP licensing powerhouse, Arm is pushing deeper into production silicon, positioning its own CPU as a direct platform for next-generation data center deployments.
At launch, Arm projected roughly $1 billion in chip revenue tied to the AGI CPU. In its latest earnings update, Arm says it now expects that figure to exceed $2 billion by fiscal year 2028, effectively doubling its earlier estimate. The company attributes the upgrade to stronger-than-expected customer demand and enthusiastic market response.
Arm also highlighted growing adoption across the AI hardware and cloud ecosystem. Companies including OpenAI, Cerebras, Positron, and Rebellions are expected to integrate Arm’s AGI CPU alongside accelerator-based compute, using the CPU as a key part of broader AI systems. On the cloud side, AI providers are also lining up to secure supply, with European AI cloud provider Verda called out as one of the deployments, planning to use the AGI CPU for agentic AI orchestration.
Beyond early integrations, Arm says AGI CPU-based commercial systems are already available for ordering through major OEM partners, including ASRock, Lenovo, Quanta, and Supermicro. That availability is important for enterprises that want a faster path from evaluation to deployment, especially as AI infrastructure spending accelerates.
Arm claims customer demand for the AGI CPU now totals more than $2 billion across fiscal year 2027 and fiscal year 2028, exceeding the levels it discussed during its earlier product event.
The company says momentum in Arm-based compute is also showing up in market share. Arm reports that its CPU compute share has reached 50% among top hyperscalers, naming Amazon, Google, and NVIDIA as key examples. If that trend continues, it strengthens Arm’s argument that its architecture is becoming a default choice for modern data center and AI infrastructure—especially for organizations looking beyond traditional x86 options.
Arm frames the AGI CPU as part of a broader “compute platform” approach, where customers can adopt Arm technology through IP, compute subsystems, or complete silicon while keeping the benefits of a common architecture and software ecosystem. With total Arm CPU shipments reaching 350 billion as of 2026 and an ecosystem of more than 22 million developers, Arm is positioning itself as not only pervasive at the edge, but increasingly influential in data centers powering large-scale AI and physical-world workloads.
