Nebius and Meta have just announced a massive AI infrastructure deal worth up to $27 billion over five years.
This agreement is a clear signal of the ongoing AI 'arms race,' where tech giants are scrambling to secure immense computing power. For Meta, this isn't just about buying more servers; it's a strategic move. The company has a massive capital expenditure (capex) plan for 2026, between $115 and $135 billion, dedicated to building out its AI capabilities. Relying on a single supplier is risky, so Meta is actively pursuing a multi-sourcing strategy. We saw this with their $14.2 billion deal with CoreWeave, and this new contract with Nebius deepens that approach, ensuring they have diverse and reliable access to the next generation of AI hardware.
So, how did Nebius, a specialized 'neocloud' provider, position itself to win such a colossal deal? It's a story of careful, strategic preparation that unfolded over the past two years. First, the company underwent a crucial corporate restructuring and resolved leadership sanctions, making it a viable partner for major Western companies. Second, Nebius secured the necessary capital, raising over $4.2 billion and, critically, landing a $2 billion strategic investment from NVIDIA itself just last week. This not only provided funds but also served as a powerful endorsement. Third, they secured the physical foundation, gaining approval for a massive 1.2-gigawatt 'AI factory' in Missouri, proving they can deliver the infrastructure at the required scale.
Timing and technology are also at the heart of this deal. The contract is explicitly tied to NVIDIA's next-generation Vera Rubin platform, which is expected to become widely available in late 2026. By structuring the deal to begin deliveries in early 2027, Meta ensures it will be among the first to leverage this cutting-edge technology, which promises significant performance improvements. This foresight allows Meta to align its long-term AI roadmap with the hardware innovation cycle.
Ultimately, this partnership is a landmark event for both companies. It provides Meta with a critical supply of advanced AI compute needed to power its future products, from ad ranking to the metaverse. For Nebius, it's a company-defining contract that provides a massive revenue backlog and cements its status as an indispensable player in the AI infrastructure ecosystem.
- Hyperscaler: A large-scale cloud computing provider that can offer services to millions of customers, such as Amazon Web Services (AWS), Google Cloud, Microsoft Azure, and Meta.
- Capex (Capital Expenditure): Funds used by a company to acquire, upgrade, and maintain physical assets such as property, industrial buildings, or equipment.
- NVIDIA Rubin Platform: NVIDIA's next-generation GPU architecture designed for AI and high-performance computing, set to succeed the Blackwell platform.
