On March 24 (local time), Arm's CEO announced the launch of its first self-developed data center chip, the Arm AGI CPU, expanding its computing platform into the realm of physical chips and targeting the data center market, which is expected to exceed $1 trillion by 2030. As the AI industry enters the stage of normalized agent reasoning in 2026, CPUs have become a critical component of modern infrastructure. Arm is committed to maximizing the physical efficiency of the AGI CPU to capture hardware premiums from traditional customers. Arm's move has triggered two major changes in the industry chain: first, it has formed an ecological alliance that exceeds expectations, leading to a power struggle with NVIDIA over control of underlying computing power; second, it has sparked backlash within the Arm ecosystem, providing momentum for the development of the RISC-V architecture. This marks the spread of AI computing power competition to the level of underlying scheduling architectures, with liquid cooling temperature control needs driven by high power consumption and the market for low-cost AI chips at the edge becoming new hotspots for capital pursuit.
