‘A Defining Moment’: Arm Enters the AI Chip Market With New AGI Processor

After years of licensing its designs to other chipmakers, Arm Holdings is finally taking matters into its own hands and building its own processor.

The company unveiled the AGI CPU, a data center chip designed specifically for artificial intelligence, marking a significant new chapter for the UK-based tech giant. CEO Rene Haas called the unveiling a “defining moment” for Arm, which has spent decades operating as a neutral platform providing intellectual property for others to build physical chips.

The company projects that the AGI CPU alone could generate $15 billion in annual revenue within five years, helping push total projected revenue to $25 billion by 2031.

Built for agentic AI

The AGI CPU is designed to handle what’s known in the tech biz as agentic AI, a class of artificial intelligence that can act on behalf of users with minimal oversight, rather than simply responding to queries like chatbots. This emerging AI category is putting new pressure on CPUs, which handle general-purpose computing tasks that GPUs aren’t well-suited for. 

Arm’s entry into producing its own CPUs positions it as both a partner and a competitor to long-established players like Intel and AMD, as well as to some of its traditional licensees, such as Nvidia, Google, and Amazon.

Meta Platforms is Arm’s first official customer for the AGI CPU, with early customers including OpenAI, Cloudflare, SAP, SK Telecom, and Cerebras. The social media giant is rapidly expanding its AI infrastructure, with plans to invest up to $135 billion in capital expenditures this year.

Performance and efficiency

The processor has been optimized for power efficiency: up to 64 AGI CPUs, totaling roughly 8,700 cores, can fit in a single air-cooled rack, offering twice the performance per watt of comparable x86-based servers.

Arm is working with Taiwan Semiconductor Manufacturing Co. (TSMC) to fabricate the chips using its 3-nanometer process. The AGI CPU is made from two distinct silicon pieces that function as a single high-performance unit and is slated for volume production later this year.

At Arm’s Austin, Texas campus, the company invested $71 million in new labs and staffed a team of more than 1,000 engineers working across design, testing, and validation to bring the CPU from concept to reality. 

Transforming Arm’s business model

Historically, Arm’s business has relied on royalties from licensing its designs, a model with exceptionally high margins.

By producing its own CPUs, Arm may see margins shrink, but it also allows for far higher revenue. Haas expects the AGI CPU to maintain an operating margin above 30%, while the traditional licensing business could double over five years.

Arm will now compete, in some sense, with the very companies it has long served, including Nvidia, Google, and Amazon, while providing new choices for data center operators looking to scale AI infrastructure efficiently.

CPU demand resurgence

The chip’s introduction also comes at a time when CPUs are poised to shine in AI workloads.

While GPUs have dominated AI training, agentic AI and cloud orchestration tasks are increasingly bottlenecked by CPU performance. Demand for these general-purpose processors is rising, and Arm hopes its energy-efficient design will provide an alternative for companies that don’t have the resources to build custom in-house chips.

Arm’s stock rose as much as 7.5% in after-hours trading following the announcement, reflecting investor optimism about the company’s new direction. The company emphasized that the AGI CPU is “available to the whole world,” suggesting ambitions to become a major supplier of AI infrastructure chips, alongside Intel, AMD, and Nvidia.

Looking to the future

With the AGI CPU, Arm is staking its claim in the rapidly growing AI infrastructure market. By offering high-performance, competitive pricing, and broad accessibility, the company hopes to redefine both the CPU landscape and its place in the tech ecosystem.

Also read: Nvidia’s AI laptop chip push includes an Arm-based path for thinner Windows systems built around integrated AI performance.