The five-year agreement centers on providing large-scale AI computing capacity to Meta. Nebius will deliver at least $12 billion in dedicated infrastructure across multiple locations, with deployment expected to begin in 2027.

In addition, Meta has committed to purchasing up to $15 billion in extra capacity over the same period if it is not used by other customers. This brings the total potential value of the deal to $27 billion, underlining the scale at which AI infrastructure is now being contracted.

The infrastructure will be built on next-generation systems powered by Nvidia technology, reflecting the increasingly tight integration between hardware providers, cloud platforms, and AI developers.

This deal is part of a broader trend in which large technology companies are securing long-term access to compute capacity. As AI models grow in size and complexity, access to data centers, GPUs, and energy is becoming a strategic priority. Specialized providers like Nebius are emerging as key players by focusing exclusively on AI workloads rather than general-purpose cloud services.

For Nebius, the agreement represents a major step in scaling its AI cloud business. It builds on earlier multi-billion deals with companies such as Microsoft and expands an existing partnership with Meta.

For Amsterdam and Europe, the implications are significant. The region is increasingly part of the global infrastructure layer that enables AI development. As investments grow and partnerships deepen, the question is no longer whether Europe participates in AI, but how it positions itself within the value chain.

This deal signals that Amsterdam is not only a place where AI is developed and applied, but also where the capacity behind it is being built at scale.