The era of intelligent embodiment of AI will ignite enormous demand for CPUs. Bernstein significantly raises the target price of Intel Corporation (INTC.US) and Arm (ARM.US).

As the artificial intelligence (AI) industry accelerates from chat Siasun Robot&Automation towards agentic AI, Wall Street has begun to reassess the long-term growth space of the server CPU market. A recent research report released by the well-known investment bank Bernstein stated that the agentic AI era will significantly increase the demand for CPUs in data centers. Therefore, the global server CPU market size (TAM) forecast for 2030 has been raised from $137 billion to $223 billion, and the target prices for AMD, Arm, and Intel Corporation have been adjusted accordingly. In particular, Bernstein raised the target price for Arm from $300 to $500, maintaining an "outperform" rating; raised the target price for AMD from $525 to $600, also maintaining an "outperform" rating; while keeping a "market perform" rating for Intel Corporation, but raised the target price from $65 to $100. At the same time, SoftBank Group, which owns a stake in Arm, also had its target price raised to 11,200 yen, higher than the previous 8,200 yen. The importance of CPUs in the transition from AI "chat Siasun Robot&Automation" to "agentic AI" is significantly increasing The Bernstein analyst team led by David Dai pointed out that the AI industry is currently transitioning from the "generative AI 1.0 era" to the "generative AI 2.0 era". If chat Siasun Robot&Automation represented AI 1.0, then agentic AI, which can autonomously plan tasks, utilize tools, and execute complex workflows, represents the direction of AI 2.0 development. The analyst stated that agentic AI requires a significant amount of task coordination and execution, with a much higher dependency on CPU resources compared to traditional chat Siasun Robot&Automation. Bernstein stated: "Agentic AI involves highly autonomous task orchestration and execution, therefore, compared to GPUs, there will be a significant increase in CPU workload." The report shows that the traditional AI data center's CPU-to-GPU ratio is typically around 1:4 or even 1:8. However, with the rapid proliferation of agentic AI, this ratio is expected to increase to 1:1 or even higher levels. In other words, for every additional GPU, more CPU resources may be required in the future. This also means that the CPU industry chain that has been overshadowed by GPUs in recent years is expected to become an important beneficiary of AI investment again. Server CPU market size in 2030 may grow sixfold compared to 2025 Based on the above assessment, Bernstein raised its forecast for the global server CPU market size in 2030 to $223 billion. In comparison, the estimated global server CPU market size in 2025 is only $37 billion. This means that the market size is expected to grow approximately sixfold in the next five years. The forecast is based on the core assumption that by 2030, global AI data center capital expenditure will reach $3.5 trillion and the CPU-to-GPU ratio in AI inference scenarios will reach 1:1. The analyst pointed out that the previously predicted $137 billion market size has now become a pessimistic scenario. Under this assumption, global AI infrastructure investment is estimated to be around $3 trillion, and the CPU-to-GPU ratio is maintained at 1:2. In an optimistic scenario, if global AI data center investment reaches $4 trillion and the CPU-to-GPU ratio increases to 1.5:1, the server CPU market size in 2030 could even reach $330 billion. Arm is seen as the biggest beneficiary in the agentic AI era Among all the beneficiaries, Bernstein is most optimistic about Arm. The analysts believe that the Arm architecture, with its excellent energy efficiency advantages, is naturally suited to the data center requirements of the agentic AI era. As AI systems continue to scale, reducing energy consumption has become an important consideration for data center operations, in which Arm has a clear advantage. More importantly, Arm is gradually extending from a simple IP licensing model to CPU product development, hoping to share more profits in the industry value chain. Bernstein previously estimated that Arm's related business revenue could reach $15 billion by 2030, but after raising the forecast for the server CPU market size, the organization has raised this number to $22 billion. The report stated: "Arm will be one of the most important structural beneficiaries under the trend of agentic AI driving CPU revival." AMD and Intel Corporation also benefit, but AMD has a more obvious advantage. Bernstein also believes that AMD and Intel Corporation will benefit from the continuous growth of data center server demand. However, from a product competitiveness perspective, AMD still maintains a leading position. The analyst stated that AMD's current server product portfolio remains superior to its competitors in terms of performance and energy efficiency, and is therefore expected to continue to expand its market share. Since Bernstein's previous performance forecasts for AMD adequately considered the growth factors in the server market, the adjustment in profit forecasts this time is relatively limited. In contrast, Intel Corporation's future performance expectations have been significantly raised, resulting in a more pronounced increase in target price.