Arm Holdings, the British chip design company that has licensed its architecture to Apple, Nvidia, Amazon, and Google for over 35 years, is now competing directly with those same customers by manufacturing its own processors. The company unveiled the AGI CPU (Arm's first in-house chip designed for artificial general intelligence workloads) on Tuesday, marking a fundamental shift in its business model from pure licensing to actual silicon production .

Meta, the social media giant building massive AI data centers across Louisiana, Ohio, and Indiana, is the first official customer. Seven other committed customers include OpenAI, Cloudflare, and SAP, with about 50 partners signaling support ahead of the launch . This is a major win for Arm, which has spent $71 million and about 18 months building three new lab rooms at its Austin, Texas campus to prepare the chip for full production later this year .

Why Is Arm Suddenly Making Its Own Chips?

For decades, Arm's business model was simple: design the instruction sets that tell processors how to work, license those designs to chipmakers, and collect royalties on every processor sold. That strategy made Arm incredibly profitable without the risk of manufacturing. So why change now?

The answer lies in a quiet crisis nobody expected. While graphics processing units (GPUs) like Nvidia's have dominated AI training and inference, something unexpected happened as AI systems became more complex. CPUs, the general-purpose processors that handle sequential tasks, became the bottleneck. Agentic AI, which uses multiple AI agents working together, requires enormous amounts of general computing power to move data between systems and coordinate tasks .

"In today's world, you really only have a couple of players. This adds yet another player to the ecosystem for us," said Paul Saab, Meta software engineer who helped with the Arm chip project since its start in 2023.

Paul Saab, Software Engineer at Meta

Nvidia itself acknowledged this shift at its recent GTC conference, where CEO Jensen Huang unveiled an entire rack filled only with Vera CPUs. The CPU market growth rate could exceed GPU growth by 2028, according to Futurum Group, which calls it a "quiet supply crisis" .

What Makes Arm's New CPU Different From Traditional Server Chips?

Most servers today run on x86 architecture, the "tried and true" standard used by Intel and AMD for decades. X86 can run pretty much anything, but it comes with a cost: it's power-hungry and less efficient than alternatives .

Arm's AGI CPU takes a different approach. The architecture is "super efficient" because the designs are more customizable, allowing engineers to build exactly what they need without unnecessary components . The result is striking: up to 64 of the new CPUs, totaling about 8,700 cores, can fit in a single air-cooled rack .

"You can get two times the performance-per-watt than you can from an x86 rack. That means twice as much performance in the same footprint, in the same power," explained Mohamed Awad, Arm's cloud AI head.

Mohamed Awad, Cloud AI Head at Arm

For data center operators like Meta, this efficiency matters enormously. Wattage is a scarce resource in large-scale AI infrastructure. If a CPU delivers better performance-per-watt, it frees up power for other critical systems like cooling, networking, and storage .

How Does This Shift Affect the Broader Chip Industry?

Arm's move into manufacturing represents a calculated risk. The company is entering direct competition with its own customers, including Amazon (which makes Graviton processors based on Arm architecture), Google, and Microsoft. Yet industry leaders appear to support the move. Jensen Huang from Nvidia appeared in a recorded statement congratulating Arm, and top executives from Google, Amazon, Microsoft, Oracle, Broadcom, Micron, Samsung, SK Hynix, and Marvell also signaled support .

The reason for this surprising support is simple: Arm's existence enables these companies to create their own processors. Without Arm's architecture, companies like Amazon, Google, and Microsoft wouldn't be able to design custom chips tailored to their specific needs .

• Market Size: The CPU market represents a $1 trillion opportunity, and Arm's entry signals that partners recognize this as beneficial for the entire industry .
• Manufacturing Location: Arm currently manufactures the AGI CPU at Taiwan Semiconductor Manufacturing Company's (TSMC) 3-nanometer fabrication plants in Taiwan, though the company has expressed interest in using TSMC's new Arizona facility once it becomes available .
• Customer Diversity: Meta's adoption is particularly significant because the company is building multiple gigawatts of AI data center capacity and plans to spend up to $135 billion on capital expenditures this year, meaning Arm could capture a substantial portion of that spending .

Steps to Understanding Arm's Competitive Advantage in Data Centers

• Performance-Per-Watt Efficiency: The AGI CPU delivers twice the performance-per-watt compared to traditional x86 processors, meaning data centers can run more AI workloads using the same amount of electricity .
• Customization Flexibility: Unlike x86 architecture, which is designed to run virtually any software, Arm's approach allows engineers to optimize for specific workloads like AI inference, reducing unnecessary complexity and power consumption .
• Supply Chain Diversification: Meta's adoption of Arm CPUs alongside chips from Nvidia and AMD reduces dependence on any single supplier, providing more flexibility in software development and procurement .
• Dense Rack Configuration: Fitting 64 CPUs with 8,700 total cores in a single air-cooled rack enables data center operators to maximize computing power in physically constrained spaces .

Chip analyst Patrick Moorhead estimates that if Arm captures just 5% of Meta's $115 to $135 billion annual capital expenditure on data centers, it would be "a game changer on the top line" for the company . This suggests the financial stakes are enormous, not just for Arm but for the entire AI infrastructure ecosystem.

The AGI CPU represents more than just a new product. It signals that the AI boom is maturing beyond GPU dominance and entering a phase where efficiency, customization, and supply chain diversity matter as much as raw performance. For Meta, which fell behind on AI development after its Llama 4 model underperformed last year, securing multiple sources of high-performance CPUs is essential to catching up . For Arm, it's a transformation from a licensing company into a direct competitor, betting that its architectural expertise can translate into manufacturing success.