Podcast Episode
The agreement represents a nine-fold increase from the initial three billion dollar contract the two companies signed in November twenty twenty-five, and marks one of the largest external infrastructure commitments Meta has ever made.
Meta has also committed to purchasing up to fifteen billion dollars in available compute capacity across certain upcoming Nebius clusters over the five-year period. Nebius intends to sell this capacity to third-party customers of its AI cloud business, with remaining capacity purchased by Meta, effectively giving the neocloud company a safety net for its expanding data centre fleet.
In September twenty twenty-five, Nebius signed a deal worth up to nineteen point four billion dollars with Microsoft to provide GPU infrastructure. Just days ago, Nvidia announced a two billion dollar strategic investment in Nebius, aimed at deploying more than five gigawatts of AI data centre capacity by twenty thirty.
Nebius Signs Massive Twenty-Seven Billion Dollar AI Infrastructure Deal With Meta
March 16, 2026
0:00
3:35
Nebius Group has signed a five-year AI infrastructure deal with Meta worth up to twenty-seven billion dollars, making it one of the largest compute-procurement contracts in history. The deal includes twelve billion in dedicated GPU capacity and fifteen billion in additional compute resources, with deployments beginning in early twenty twenty-seven.
A Record-Breaking AI Infrastructure Partnership
Meta Platforms has signed a landmark five-year AI infrastructure agreement with Nebius Group valued at up to twenty-seven billion dollars, dramatically expanding the relationship between the social media giant and the Amsterdam-based neocloud provider.The agreement represents a nine-fold increase from the initial three billion dollar contract the two companies signed in November twenty twenty-five, and marks one of the largest external infrastructure commitments Meta has ever made.
How the Deal Works
Under the agreement, Nebius will provide twelve billion dollars in dedicated GPU-based capacity across multiple locations, with deployments beginning in early twenty twenty-seven. The infrastructure will be built on one of the first large-scale deployments of Nvidia's Vera Rubin platform, its next-generation chip architecture.Meta has also committed to purchasing up to fifteen billion dollars in available compute capacity across certain upcoming Nebius clusters over the five-year period. Nebius intends to sell this capacity to third-party customers of its AI cloud business, with remaining capacity purchased by Meta, effectively giving the neocloud company a safety net for its expanding data centre fleet.
Nebius: From Yandex Spinoff to AI Powerhouse
Nebius has rapidly emerged as one of the fastest-growing companies in the AI infrastructure sector. The company, which was previously the Dutch holding company for Russian internet giant Yandex before divesting its Russian assets in twenty twenty-four, has assembled an impressive roster of major customers in a remarkably short time.In September twenty twenty-five, Nebius signed a deal worth up to nineteen point four billion dollars with Microsoft to provide GPU infrastructure. Just days ago, Nvidia announced a two billion dollar strategic investment in Nebius, aimed at deploying more than five gigawatts of AI data centre capacity by twenty thirty.
The Bigger Picture
The expanded deal underscores the intensifying competition among technology companies to secure computing capacity for AI development. Meta has outlined plans to spend between one hundred and fifteen billion and one hundred and thirty-five billion dollars on capital expenditure in twenty twenty-six. Nebius itself is targeting three to three point four billion dollars in revenue for twenty twenty-six, up from five hundred and thirty million last year, and plans capital expenditure of sixteen to twenty billion dollars as it scales from seven data centre sites to sixteen.Published March 16, 2026 at 2:12pm