Renowned chip and software firm Arm Holdings is embarking on the production of its proprietary processors, after nearly 36 years of exclusively licensing its designs to companies such as Nvidia and Apple.
During a gathering on Tuesday in San Francisco, the firm unveiled the Arm AGI CPU, a processor prepared for mass manufacture engineered to execute inference operations within an artificial intelligence data facility. The British-headquartered enterprise crafted the processor leveraging its Arm Neoverse family of CPU IP cores and through a collaboration with Meta.
Furthermore, Meta serves as the initial purchaser of the Arm AGI CPU, which is conceptualized to integrate seamlessly with the technology firm’s training and inference accelerator. Arm additionally recognizes OpenAI, Cerebras, and Cloudflare, among other entities, as initial collaborators.
Arm’s shift towards producing its proprietary silicon had been expected for a while. The firm began crafting these processors as early as 2023, as per reports from CNBC, with the chips already available for procurement.
TechCrunch contacted Arm seeking further details concerning the schedule of the processor’s creation and launch.
Though perhaps foreseeable, this step marks a significant departure from Arm’s long-standing practice of solely granting permission to use its blueprints to alternative semiconductor manufacturers. The firm, predominantly owned by the Japanese conglomerate Softbank Group, will now contend with numerous of its associates.
The circumstance that Arm is manufacturing a CPU, rather than a GPU, is likewise noteworthy. GPUs, or graphics processing units, have garnered considerable interest since they are employed for training and executing AI models. CPUs constitute an equally crucial component of a data center’s infrastructure.
Techcrunch event
San Francisco, CA
|
October 13-15, 2026
In its argument advocating for CPUs, Arm highlights that these processors oversee myriad dispersed operations, encompassing memory and storage administration, workload orchestration, and data transit across diverse systems. The CPU has evolved into the “critical component of contemporary infrastructure — tasked with ensuring distributed AI systems function optimally on a large scale,” as stated by the company.
This imposes fresh requirements on CPUs, necessitating a progression in processor design, Arm declared.
Obtaining CPUs is also growing more challenging.
During March, Intel and AMD informed their Chinese clientele that delivery periods for their merchandise would extend due to a scarcity of CPUs, Reuters initially disclosed this. Concurrently, the cost of computers has begun to escalate amidst the increasing scarcity.
{content}
Source: {feed_title}