Editorial illustration for Arm's first AGI CPU, up to 136 cores, to power Meta AI datacenters this year
Arm's AGI CPU Supercharges Meta's AI Data Centers
Arm's first AGI CPU, up to 136 cores, to power Meta AI datacenters this year
The tech world has been watching Arm’s push into high‑density AI silicon for months. While the company touts an AGI‑focused processor that can house up to 136 cores on a single die, the real question is where that muscle will end up. Meta’s data centers, already humming with massive workloads, are slated to receive the new chips later this year—a move that could reshape power efficiency calculations across the industry.
If the claim of double the performance‑per‑watt versus traditional x86 designs holds up, operators might finally see a tangible break from the thermal and cost constraints that have limited scaling. But the rollout isn’t just about raw numbers; it’s a test of how quickly a novel architecture can be woven into an existing, massive compute fabric. Here’s the thing: the partnership signals more than a hardware swap—it hints at a broader shift toward specialized, energy‑aware AI engines.
**Arm's first CPU ever will plug into Meta's AI datacenters later this year**
Arm's first CPU ever will plug into Meta's AI datacenters later this year The Arm AGI CPU can have up to 136 cores per CPU, claiming it has double the performance per watt of x86 chips. The Arm AGI CPU can have up to 136 cores per CPU, claiming it has double the performance per watt of x86 chips. Meta says it's both the lead partner and co-developer, and plans to work on "multiple generations" of the datacenter CPUs, for use along with hardware from other vendors like Nvidia and AMD.
Arm customers like Amazon AWS, Microsoft, Google, Marvell, Nvidia, Samsung, and others included congratulatory notes with the announcement. However, Qualcomm, which said it had achieved "complete victory" over Arm with a court ruling last fall in their case over the terms of licensing agreements, was not one of them. Financial terms of the deal weren't revealed, nor was the number of chips Meta plans to use from Arm, which is currently owned by Softbank.
According to Arm, its new chip runs on the Neoverse platform used by AWS Graviton, Nvidia Vera, Microsoft, and other AI chips, with up to 136 cores per CPU and 64 CPUs per air-cooled server rack. It says the AGI CPU can get twice as much performance per watt as traditional x86 CPUs while reducing memory bottlenecks, taking advantage of the design's long-running efficiency advantages. Other customers lined up for Arm's chip include Cerebras, Cloudflare, F5, OpenAI, Positron, Rebellions, SAP, and SK Telecom.
Arm cloud AI head Mohamed Awad told CNBC that its aim is to be an option for companies that can't afford to make their own in-house processor.
Will the Arm AGI CPU live up to its promises? The chip, Arm’s first self‑produced silicon, is slated to arrive in Meta’s AI datacenters before year‑end. Designed for inference, it targets workloads where AI agents spawn numerous parallel tasks.
With up to 136 cores, Arm claims the processor delivers roughly twice the performance per watt of comparable x86 designs. Meta, which has faced difficulties bringing its own AI silicon to market, will be the inaugural customer. Yet the claim of double efficiency lacks independent verification, and real‑world testing at scale has not yet been disclosed.
If the cores can be effectively utilized across diverse models, the architecture could offer a notable alternative to existing solutions. Conversely, the complexity of managing so many cores may introduce scheduling overheads that erode the advertised gains. The partnership marks a shift for Arm, moving from pure licensing to manufacturing, but whether the AGI CPU will become a mainstay in large‑scale AI infrastructure remains uncertain.
Further Reading
- Meta and Arm Co-Develop First AI-Era Data Center CPU - TechBuzz
- Meta Partners With Arm to Develop New Class of Data Center Silicon - Meta
- Arm rolls its own 136-core AGI CPU to chase AI hype train - The Register
- Arm expands compute platform to silicon products in ... - Arm Newsroom
- Arm is releasing its first in-house chip in its 35-year history - TechCrunch
Related Reading
Browse all Meta news →
Business & Startups OpenAI, a Series F San Francisco startup founded in 2015 by eight pioneers
Business & Startups Terminal-Bench 2.0 launches with Harbor, testing any container-installable agent
Business & Startups Zuckerberg Unveils Meta Compute to Build Global AI Infrastructure
Related Topic Meta's DreamGym boosts AI agent success by 30% over baseline methods
LLMs & Generative AI
Related Topic Study finds condensed metaphors and rhythmic framing can evade safety filters
LLMs & Generative AI
From Archives Moda assesses migration of Design Agent from custom LangGraph loop
From Archives Nvidia CEO clarifies DLSS 5 concerns in lengthy Lex Fridman interview
From Archives Meta engineer's use of internal AI agent triggers serious security incident
Open Source
From Archives Nvidia unveils Vera Rubin platform OpenAI, Anthropic, Meta; adds NemoClaw stack
AI Tools & Apps
Common Questions Answered
How many cores will Arm's new AGI CPU have in Meta's data centers?
Arm's new AGI CPU will feature up to 136 cores on a single die, representing a significant leap in processor density. This high-core-count design aims to dramatically improve performance efficiency for AI workloads in Meta's data centers.
What performance advantage does Arm claim for its new AGI CPU compared to x86 chips?
Arm claims its new AGI CPU delivers double the performance per watt compared to traditional x86 chip designs. This efficiency improvement could potentially revolutionize power consumption and computational capabilities in AI data center infrastructure.
What is Meta's role in the development of Arm's new AGI CPU?
Meta is both the lead partner and co-developer of the new Arm AGI CPU, planning to work on multiple generations of datacenter CPUs. The company will be the inaugural customer, integrating the chip into its AI data centers later this year.