Arm, one of The world’s leading chip design companies announced Tuesday that they are producing their own semiconductors. The move marks a departure from the long-standing model of licensing intellectual property to companies that produce and sell chips themselves. Speaking to an audience in San Francisco, Arm CEO Rene Haas outlined his pitch for how the up-to-date Arm processor could benefit the tech industry and why it’s the right time for the company to step out of its lane and go head-to-head with other chipmakers.
Arm’s internal chip efforts have been talked about for years. Now, as artificial intelligence spreads across the economy and demand for computing resources skyrockets, Arm is trying to capture a slice of the market for central processing units (CPUs) optimized to handle AI workloads.
The up-to-date chip is called the Arm AGI processor, a nod to artificial general intelligence, an oft-referenced but still hypothetical form of artificial intelligence that can match human performance in various domains. It is designed to interface with other chips in high-performance servers in data centers and to handle agent-based AI tasks. The chip is manufactured by Taiwan Semiconductor Manufacturing Corporation, the world’s leading semiconductor foundry, and is built using TSMC’s 3nm process.
During the chip presentation, Arm executives highlighted the company’s history in designing energy-efficient chips and said its up-to-date AGI processor will be “the most efficient agent processor on the market.” Compared to competitors such as the latest x86 chips from Intel and AMD, Arm says the chip will provide better performance per watt, or the amount of energy a computer uses, and could save customers billions of dollars in electricity costs.
The first major customer of the up-to-date Arma chip is Meta, which the company says has received samples of the processor. OpenAI, SAP, Cerebras and Cloudflare, as well as Korean tech companies SK Telecom and Rebellions have also agreed to buy the chip. Arm expects the AGI processor to reach “full production availability” in the second half of this year.
Nvidia CEO Jensen Huang, Amazon senior vice president and distinguished engineer James Hamilton, and Google AI infrastructure chief Amin Vahdat appeared in videotaped testimonials praising the up-to-date Arm hardware. None has decided to purchase it, but all three tech giants already operate Arma designs in their own processors.
Arm’s history dates back to the delayed 1970s, when the company was known as Acorn and manufactured microprocessors. In the 1990s, the company changed its name to ARM (Advanced RISC Machines), and its then-CEO began licensing the company’s chip designs to other companies. Arm, which has since abandoned the all-caps “ARM” brand, has seen business boom during the mobile revolution. By 2010, many of the world’s largest technology companies, including Apple, Nvidia, Microsoft, Amazon, Samsung, and Tesla, relied on this technology.
Arm eagerly attended the press event to demonstrate that it has support from bold names in the tech industry. While the company is mainly targeting chipmakers such as AMD and Intel that build processors based on different architectures, it risks potentially alienating some of its longtime partners by releasing its own chips. Nvidia, which primarily produces graphics processors, also includes ARM-based processors in its rack systems. Earlier this year, Nvidia announced that it would sell standalone processors for the first time. Meta was one of its first buyers.
Ben Bajarin, CEO and principal analyst at research firm Inventive Strategies, says Arm may be viewed more as a competitor than a partner as its strategy evolves. Right now, Arm is rolling out a streamlined processor with a relatively tiny number of cores – the chip’s built-in processing units – designed specifically to run AI agents, Bajarin points out. Over time, Arm may expand into more general-purpose processors, while AMD and Intel will develop chips tailored to agent-based AI. This would put companies in more direct competition with each other.