A recent report from Citi suggests the server CPU market is on the verge of a significant transformation, forecasting it to reach $131.5 billion by 2030.
The driving force behind this growth is a new category called 'agentic CPUs.' In the past, the AI conversation was dominated by GPUs, the powerful chips needed for training massive AI models. However, as AI evolves into autonomous 'agents' that can perform complex tasks on their own, a new need has emerged: a central 'brain' to coordinate everything. This is the new role for the CPU—acting as an orchestrator or 'control plane' that directs GPUs, memory, and thousands of simultaneous AI agent tasks. The AI story is expanding from just training models to deploying intelligent systems, and CPUs are at the heart of this deployment.
So, why is this happening now? The shift is supported by a few key developments. First, the necessary hardware is becoming available. Arm recently launched its AGI CPU, a chip specifically designed for this orchestration role. This launch transformed the concept of an agentic CPU from an abstract idea into a tangible product that companies can build systems around. It signaled that the industry is specializing its silicon for the agentic era.
Second, the competitive landscape is already in motion. AMD has been steadily gaining server market share from its rival Intel, proving that the market is dynamic and hungry for innovation. This competitive pressure encourages faster development across the board. Furthermore, NVIDIA is structurally linking CPU and GPU demand with its Grace-Blackwell superchip, which tightly integrates both processors. This means that as companies buy more GPUs for AI, they will inherently buy more high-performance CPUs as well.
Finally, the world's largest cloud providers—Amazon (AWS), Google, and Microsoft—are validating this trend by developing their own custom Arm-based CPUs (Graviton, Axion, and Cobalt). This move by the hyperscalers not only proves the efficiency of the Arm architecture at a massive scale but also cements the idea of a multi-vendor CPU market where customized solutions thrive. Citi's forecast isn't a guess; it's a reflection of these powerful, converging trends that are reshaping the semiconductor industry.
- Glossary -
- Agentic AI: AI systems that can act autonomously to achieve goals by breaking down complex tasks into smaller steps and executing them.
- TAM (Total Addressable Market): The total potential revenue if a product or service achieved 100% market share in a specific market.
- Orchestration: In computing, the process of coordinating and managing multiple systems, applications, and services to work together harmoniously.