ARM Holdings reported a record-breaking fourth quarter for fiscal year 2026, driven by surging demand for its new AGI CPU product, which saw customer commitments double to $20 billion within just six weeks of launch. This surge is attributed to increasing computational requirements from AI agents that demand enhanced underlying processing coordination.

After U.S. markets closed on May 6, ARM announced its fiscal Q4 revenue rose 20% year-over-year to $1.49 billion, exceeding the midpoint of previous guidance. Full-year revenue increased 23% to $4.92 billion, marking the third consecutive year of over 20% revenue growth.

During the earnings call, CEO Rene Haas revealed that customer demand for the ARM AGI CPU has doubled since its introduction six weeks ago, surpassing $20 billion—far exceeding the initial $10 billion supply target. However, supply chain bottlenecks remain a key constraint, leading the company to maintain its $1 billion annual revenue guidance for the product while actively expanding wafer, packaging, and testing capacity.

Data center royalty revenue doubled year-over-year, with Haas affirming expectations for another doubling and reiterating the long-term goal of reaching $25 billion in revenue by fiscal 2031.

**AGI Chip Demand Soars: Orders Double to $20 Billion in Six Weeks**

A key highlight of the earnings report was the ARM AGI CPU, unveiled in late March at the "Arm Everywhere" event. ARM's strategic shift from pure IP licensing to custom chip development has progressed faster than management anticipated.

Haas expressed strong confidence in the product, stating, "Customer response to the ARM AGI CPU has been tremendous. We now have over $20 billion in customer demand spanning fiscal 2027 and 2028—more than double the figure we announced at launch."

Despite soaring demand, supply limitations led CFO Jason Child to maintain the $10 billion revenue target, confirming that revenue from initial production shipments will be recognized in Q4 of fiscal 2027.

**AI Agents Drive Trillion-Dollar Market: Core Count Matters More Than Chip Count**

A major topic of discussion was the growth potential of CPUs in an AI landscape dominated by GPUs. Haas argued that as AI evolves from human-driven queries to continuous, agent-driven workloads, the role of the CPU expands significantly.

"These agent workloads require CPUs to coordinate tasks, move data, manage memory, enforce security, and orchestrate work around accelerators," Haas explained. "As AI agents scale, data centers will need more than four times today’s CPU capacity, creating a data center CPU market opportunity exceeding $100 billion by 2030."

Responding to a competitor’s estimate of a $1.2 trillion market, Haas acknowledged the potential for a larger total addressable market. He emphasized that core count, rather than the number of chips, is the more accurate metric for evaluating market dynamics.

Haas noted that while accelerators like NVIDIA’s Blackwell and Rubin are constrained by reticle size limits, CPUs can scale by increasing core counts. "The ARM AGI CPU already has 136 cores, and Vera has 88. Doubling or even quadrupling core counts in the coming years is entirely feasible," he said. "If a chip goes from 136 to 500 cores, the chip count ratio may not change much, but the core count ratio will shift fundamentally."

**"Largest CPU Market Share by 2030 Will Belong to ARM"**

When discussing competition with AMD and Intel, ARM expressed strong confidence in its ecosystem expansion. Leading hyperscalers, including NVIDIA’s Vera, Google’s Axion, Amazon AWS’s Graviton, and Microsoft’s Cobalt, have already adopted custom CPUs based on the ARM architecture.

Haas stated, "I expect that over time, the vast majority of the market share for Trainium, TPU, and NVIDIA accelerators will be ARM-based. I am confident that by 2030, the largest share of the CPU market by architecture will belong to ARM."

This ecosystem adoption is driven by significant energy efficiency gains. Haas cited Google’s recent TPU 8t and 8i releases, which replaced x86 host processors with custom ARM Axion CPUs, achieving performance improvements with 50% lower power consumption and an 80% overall architectural enhancement compared to prior x86 solutions.

**Dual Growth Engines Support Long-Term $25 Billion Revenue Target**

ARM is undergoing a significant business model evolution, expanding from selling IP and Compute Subsystems (CSS) to directly offering chips. This creates a unique barrier through a unified compute platform and software ecosystem.

Haas emphasized that the strategy has received support from over 50 industry leaders, including AWS, Microsoft, Google, NVIDIA, and TSMC. "The more software written and optimized for ARM, the stronger the entire ecosystem becomes," he noted.

This dual-engine model provides strong revenue visibility. The company expects AGI CPU revenue to reach $15 billion by fiscal 2031, while IP revenue doubles to $10 billion, totaling $25 billion in revenue and translating to earnings per share exceeding $9.

For fiscal 2027, CFO Jason Child guided Q1 revenue to approximately $1.26 billion, representing roughly 20% year-over-year growth, with both royalty and licensing revenue expected to maintain around 20% growth for the full year.