Arm Holdings launches Arm AGI CPU for AI data center workloads, positioning the architecture as a high-efficiency alternative to x86 systems amid expanding adoption by cloud giants including Google and NVIDIA.

Arm Holdings introduced Arm AGI CPU, a processor designed for agentic AI workloads (AI agents that operate autonomously). The company states the architecture offers significantly stronger rack-level performance and improved infrastructure economics compared to conventional x86 systems, and claims it can help customers lower capital spending while scaling AI workloads across cloud and enterprise environments.

Google is integrating custom Axion CPUs into upcoming TPU systems and stated its next-generation TPU infrastructure will replace legacy x86 host processors with Arm-based designs. NVIDIA emphasized deeper integration of Arm-based CPUs into next-generation AI systems and introduced its Vera CPU platform. Companies including AWS, Microsoft, Oracle, Samsung, Micron, and SK Hynix are supporting broader adoption of the Arm ecosystem.

Arm stated that customer demand for the AGI CPU has already surpassed early launch expectations, supported by cloud operators, telecom infrastructure providers, and enterprise AI customers. The company also noted that Arm-based compute now represents roughly half of CPU deployment among major hyperscalers.