Server CPUs are one of the most overlooked AI beneficiaries as agentic inference shifts 70-80% of the workload back to CPUs. Per UBS's Timothy Arcuri, server CPU TAM grows about 5x, from $31B in 2025 to $170B by 2030, with AI CPUs rising from $7B to $120-125B and ASPs climbing. Arm gains the most share, from 15% to 40-45% of units; AMD's high-core-count EPYC suits multi-agent workloads; Intel benefits from the rising tide but keeps losing share. x.com 

AI agents are pushing server CPU demand far beyond current supply, making Arm, AMD and Intel CPUs one of the most attractive ways to get exposure to agentic AI. At about 10 agents per CPU, one agent for each of 100 million US knowledge workers needs 10 million CPUs, and 10 agents each for 1 billion users globally needs 1 billion, about 20x the world's 35-45 million annual server CPU output. AMD now sizes the 2030 server CPU TAM at about $220B, up from $120B, versus a roughly $26B market in 2025. … Show more

Intel and AMD face a real shift in competition as NVIDIA moves into CPUs faster and more directly than expected. On NVIDIA's earnings call, Jensen Huang said its new Vera CPU platform will contribute $20 billion of standalone revenue this year. The pressure on Intel and AMD is not just fund rotation but a genuine change in the competitive landscape. x.com 

Agentic AI demand should drive AMD's 2027 data-center growth well over 100% year on year. Server CPU revenue is forecast to double, with over 5 million Venice units shipping in 2027; Helios racks should ramp mostly in Q1 2027, with over 20,000 expected that year, while MI450/455X volume remains limited by end-2026. Projected 2027 sales are $103bn, with 2027 EPS revised up 15% and a new $720 target at 40x 2027 earnings. x.com 

Intel stands to benefit because agentic AI has left the whole CPU category short, not just handed share to AMD. Agents spend most of their time outside the model call, on tool calls, browsers, sandboxes, retrieval and orchestration, pushing CPU-to-GPU ratios from 1:8 in training and 1:4 in inference toward 1:1. Arm sees 120M CPU cores per gigawatt in agent-era data centers versus 30M before. AMD doubled its server CPU TAM to $120B by 2030, and Intel is filling only 40% of backlog. x.com 

Intel's capex raise, once seen as good news, is now a concern as Intel keeps burning cash while pushing toward nearly $100B of US capital investment this decade. AMD coming for its CPU business adds to the pressure. x.com 

US sovereign AI and defense demand for domestically made compute will bring Intel most, if not all, US semiconductor designers as customers. PC weakness also lets Intel shift wafers to higher-margin server CPUs, but that takes a few quarters; capacity to ship more server CPUs is the limiter. The $240 billion 2030 server CPU market forecast is a stretch case, not the base case, versus roughly $200 billion consensus. x.com  x.com  x.com  x.com  x.com 

Intel, AMD and Arm benefit because agentic AI needs far more CPUs per GPU. J.P. Morgan models the GPU-to-CPU ratio falling from 4.0x in 2024 to 1.9x by 2028, roughly doubling CPUs per GPU on top of growing GPU shipments, and that is the conservative view. TrendForce expects one or two GPUs per CPU in agentic deployments, and Intel CEO Lip-Bu Tan said some frontier labs cite four CPUs per GPU. Every orchestration step and agent sandbox runs on a CPU, and Meta gave every Muse user an always-on VM. … Show more

AI agents like Meta's Muse need far more CPUs, benefiting Intel, AMD and China's Hygon. Orchestration runs on CPUs and accounts for 50-90% of agent latency; the CPU-to-GPU ratio is converging from 1:8 toward 1:1, and CPU cores needed per gigawatt of data center could jump from 30 million to 120 million. Muse topped the US App Store and passed 2.5 million downloads in two weeks. How Meta recoups heavy users' $20-40 token costs is unclear, and agent business models may be harder than the tech. … Show more

AMD stands to gain as CPUs for agentic AI could exceed 50% of the total CPU market by 2030; @BenBajarin estimates a 62% CAGR for agentic CPUs. The risk is supply: next year ASIC makers and AMD may struggle with supply chain capacity constraints, as Nvidia locks up tier-one suppliers with hard supply agreements, which per @lokoyacap signals worsening shortages. That lockup is good for tier-two suppliers. x.com  x.com