Meta has deepened its alliance with ARM with the aim of delving into the advances in infrastructure that are planned to advance in the field of Artificial Intelligence. Among other things, with the new agreement signed, Meta, which has already been a long-time ARM customer, has confirmed that, among other planned implementations, its ranking and recommendation systems will be moved to ARM’s Neoverse platform, which has been optimized a few weeks ago to work with AI systems in the cloud.
ARM, whose GPU offering is often overshadowed by that of NVIDIA and other competitors, is in a campaign to show its advantages in low-power deployments. And Meta seems to have convinced himself of this.
Unlike what has happened with other recent agreements between large technology companies, ARM and Meta have not exchanged shares of the companies. Nor physical infrastructure. Meanwhile, NVIDIA has committed multimillion-dollar investments in AI-related companies. As a result, according to Techcrunch, it will spend $100 billion on OpenAI, as well as $1 billion on xAI, as well as Thinking Machines Lab, the startup of Mira Murati, one of the founders of OpenAI. Also, among others, in the French Mistral.
The multi-year agreement between Meta and ARM comes at the same time that Meta has accelerated its investment in the expansion of its data center network to keep up with the pace necessary to meet the demand for AI services.
One of their projects, which they have called Prometheus, with several gigawatts of power, will be connected to the grid in 2027. It is currently under construction in Ohio (United States), and will have a 200 Megawatt natural gas project coupled to cover its energy needs.
Meta is also building a data center campus, Hyperion, in Northwest Louisiana (USA). When it is completed, which is estimated to happen around 2030, although some areas may be ready sooner, it will offer five gigawatts of computing power.
Santosh Janardhan, Infrastructure Manager at Metahas pointed out that «AI is transforming the way people connect and create. Reaching an agreement with ARM allows us to efficiently scale this innovation to the more than 3 billion people who use Meta’s technologies and apps.«.
For his part, the CEO de ARM, Rene Haashas commented that «The next stage of AI will be defined by delivering efficiency at scale. By partnering with Meta we are uniting ARM’s performance per watt leadership with Meta’s AI innovation.«.
