Agentic AI is reshaping the architecture of global compute. Agents are reasoning, deciding, and acting continuously and simultaneously at global scale. This shift is driving a once-in-a-generation rethink of computing, resetting the language of capacity to tokens, petabytes, and gigawatts.

The world needs a new class of compute built for this reality, and we believe partners across the ecosystem are looking to deploy Arm technology at scale. In response, Arm has extended its compute platform to provide unique optionality to partners—IP, compute subsystems, and now production silicon.