The landscape of the CPU market is being rapidly reshuffled, accelerated by the wave of AI agents.

At NVIDIA's recent Taipei GTC conference, CEO Jensen Huang confirmed the Vera CPU is entering mass production, following the company's announcement of nearly $20 billion in revenue visibility for its CPU business this year. Huang stated directly, "In the age of AI agents, the CPU has become the bottleneck that constrains GPU performance."

Regarding this battle for CPU market share, a recent semiconductor industry report from Barclays systematically analyzes the competitive landscape and the prospects for each player. The report suggests this is a hundred-billion-dollar market being repriced, with Advanced Micro Devices (NASDAQ: AMD) as the biggest winner, Intel (NASDAQ: INTC) as the biggest loser, and NVIDIA (NASDAQ: NVDA) as the most formidable new entrant.

The CPU's New Prominence

For the past few years, the semiconductor market's focus has been on GPUs. However, over the last six months, AI applications have quietly shifted—from model training to inference and agentic AI.

This shift directly drives CPU demand. The reason is simple: GPUs excel at large-scale matrix computations, but AI agents need to execute tasks, call tools, route sub-agent traffic, and track task completion—all jobs for the CPU. The GPU "computes," while the CPU "manages."

At its "ARM Everywhere" event, ARM presented data indicating a 1GW-scale AI data center currently requires about 30 million CPU cores, a number that will grow to 120 million in the future. Based on 136 cores per ARM CPU, this equates to roughly 221,000 CPUs per GW data center. NVIDIA's CEO Jensen Huang estimates about 300,000 Rubin GPUs are needed per GW, leading to a new CPU-to-GPU ratio of approximately 1:1.4—a significant elevation in the CPU's importance compared to the GPU-dominated era.

Market Size Projections Are Rising Fast

Forecasts for the CPU market size are being revised upward rapidly.

NVIDIA: Expects its own CPU demand to reach $20 billion by 2026, with the total industry addressable market (TAM) hitting $200 billion by 2030.

ARM: Projects the AGI CPU market will exceed $100 billion by 2030.

AMD: Has raised its server CPU market forecast to $120 billion by 2030 and expects to capture over a 50% share.

After synthesizing Mercury industry data and company-specific bottom-up forecasts, analysts offer a more aggressive view: the CPU market size could approach $200 billion by 2030, exceeding AMD's own $120 billion projection. The report notes, "Market forecasts have doubled in the past six months, showing how fast this market is changing."

NVIDIA: The Most Disruptive Newcomer

NVIDIA's CPU ambitions are no secret. Previously, CFO Colette Kress stated on an earnings call that the Vera CPU opens a potential $200 billion market "previously unaddressed" by NVIDIA, with nearly $20 billion in revenue visibility this year, and that Vera is "propelling us to become a leading CPU supplier worldwide."

At the GTC conference, Huang further framed the issue: "In the AI agent era, the CPU has become the bottleneck that constrains GPU performance." The Vera CPU reportedly offers 3x better performance on SQL operations, 6x better on data processing, and about 1.8x better overall performance on common agent tools like Python and code analysis compared to x86 competitors. Anthropic and OpenAI are early adopters.

How to view this $20 billion? Analysts note that almost all of NVIDIA's CPU sales are bundled within its NVL systems (rack-scale supercomputing systems), with no independent average selling price (ASP) data point available. Reverse-engineering the $20 billion figure suggests NVIDIA's CPU ASP is significantly higher than its peers. If calculated using ASPs closer to industry peers, AMD's market share would be even nearer to its stated 50% target.

NVIDIA's entry is changing not just market share but the entire logic of system ratios. Historically, the ratio was roughly two GPUs to one CPU; now it is approaching 1:1. NVIDIA has also introduced a standalone CPU rack with 256 CPUs, with Anthropic and OpenAI as early customers.

Notably, in Barclays' modeling, NVIDIA's CPU revenue is "embedded" within NVL rack calculations and not listed separately—suggesting the market may not have fully priced in the scale of NVIDIA's CPU business.

AMD: The Primary Beneficiary

Analysts view AMD as the "most under-priced" beneficiary in this CPU battle, meaning the market has not fully reflected its potential.

AMD's advantages stem from several converging factors:

On the product side, AMD's Venice and Verano series are expected to migrate to a 2nm process first, potentially widening its performance lead. Its Chiplet architecture offers high flexibility, allowing for customization with base cores, I/O, and 25%-30% differentiated variants for different applications without a full design cycle. AMD is already shipping agent-style rack CPU systems.

On the supply side, AMD's deep ties with TSMC and the replaceability of its Chiplet platform allow for flexible allocation of wafer capacity—shifting from gaming and client businesses to data centers. Management has stated that the "narrative of supply constraints preventing near-term upside" is a "false narrative." AMD has also committed $10 billion to the Taiwan ecosystem for developing EFB (Elevated Fan-out Bridge) and panel-level packaging technologies, partnering with ASE, SPIL, and PTI to diversify future packaging needs.

Regarding market structure, AMD segments the server CPU market into three parts: 20% general-purpose enterprise CPUs (low core count); 30%-35% head node CPUs (high-frequency single-thread); and the remaining 45%-50% high-core-count CPUs for agentic AI, which may be deployed in independent racks or within data centers. AMD's product line covers all three segments.

Analysts forecast AMD's CPU revenue will reach approximately $29 billion by 2027. The CPU business alone could contribute about $19 in earnings per share (EPS) by 2030, effectively more than doubling the company's overall EPS.

Based on this, analysts have raised their price target for AMD from $500 to $665, implying a 44x multiple on projected 2027 EPS of $15.14 (up from a 37x multiple on $13.40). The "Overweight" rating is maintained.

Intel: On the Defensive

In contrast, Intel's position is seen as the most passive.

Analysts maintain an "Equal Weight" rating on Intel, raising the price target from $65 to $100 (primarily reflecting valuation repair due to higher ASPs, not market share improvement). Their assessment is blunt: "Core market share continues to erode, revenue and profitability are structurally pressured, and there is no evidence of a return to manufacturing leadership."

In the 2030 CPU market EPS contribution estimates, Intel is projected at about $1.5, compared to approximately $19 for AMD and $7 for ARM—highlighting a clear gap.

Intel's challenges are not solvable in a single product cycle: the efficiency disadvantage of the x86 architecture for AI agent workloads, persistent manufacturing process lag, and market share pressure from both AMD and NVIDIA combine into a structural predicament difficult to reverse quickly.

ARM: The Royalty Collector

ARM's proposition is relatively straightforward: it collects royalties whenever a CPU uses its architecture, regardless of the manufacturer. NVIDIA's Vera uses ARM architecture, and some of AMD's products are also migrating towards ARM.

Analysts project ARM's CPU business will contribute about $7 in EPS by 2030, raising the price target from $250 to $360, implying a 115x multiple on projected 2027 EPS of $3.13 (up from 80x). The "Overweight" rating is maintained.

The logic for the higher valuation multiple is that the expanding CPU market significantly raises the ceiling for ARM's royalty income, while ARM itself does not directly engage in market share battles, leading to a more stable benefit.

Which Player Remains Underappreciated?

The conclusion is that AMD's CPU upside is the least fully priced in among the three major players.

The reasons are threefold: First, NVIDIA's CPU revenue is partially embedded in the market's pricing of its NVL systems. Second, ARM's high valuation already reflects considerable optimism. Third, AMD's CPU share gains, process node leadership, and supply flexibility are not yet fully factored into its stock price.

A CPU market restructuring triggered by AI agents is rapidly altering the competitive dynamics of the semiconductor industry. What was once a two-horse race between Intel and AMD now sees NVIDIA as a formidable new entrant, and the market's total size is expanding at an unprecedented rate.