Millions of AI chips for new infrastructure
The Meta-Nvidia deal will see the social media giant purchase millions of artificial intelligence chips for the dozens of data centres it is building.
Meta, the parent company of Facebook, Instagram, Threads and WhatsApp, will deploy Nvidia hardware across its AI facilities, including the company’s standalone CPUs and next-generation Vera Rubin systems, under a wide-ranging agreement announced on Tuesday. Financial terms were not disclosed.
Chief executive Mark Zuckerberg said the expanded partnership supports Meta’s effort to deliver “personal superintelligence to everyone in the world”, a vision he first outlined in July.
Market reaction and spending plans
Shares in both Meta and Nvidia rose in after-hours trading following news of the deal, while Advanced Micro Devices fell by around 4%.
In January, Meta announced plans to spend up to $135 billion on artificial intelligence in 2026. Industry analyst Ben Bajarin of Creative Strategies said the agreement is “certainly worth tens of billions of dollars”, adding that a significant portion of Meta’s capital expenditure will likely flow into the expanded Nvidia partnership.
A deeper long-term partnership
The collaboration between the two Silicon Valley companies is not new. Meta has used Nvidia graphics processing units for at least a decade. However, the Meta-Nvidia deal marks a far broader technological partnership.
A key development involves standalone CPUs. Meta will become the first company to deploy Nvidia’s Grace central processing units independently in its data centres rather than pairing them with GPUs in a single server. Nvidia described this as the first large-scale standalone deployment of Grace CPUs, while next-generation Vera CPUs are scheduled for installation from 2027.
Massive US data centre investment
The multi-year agreement forms part of Meta’s broader commitment to invest $600 billion in the United States by 2028 in data centres and related infrastructure.
Meta plans to build 30 data centres, 26 of them in the US. Construction is already under way on two of the largest AI facilities: the 1-gigawatt Prometheus site in New Albany, Ohio, and the 5-gigawatt Hyperion site in Richland Parish, Louisiana.
Also read: Uber Q4 revenue up 20%, boosted by food delivery
For more videos and updates, check out our YouTube channel
