Meta’s €27 Billion Bet on European AI Infrastructure

European AI infrastructure just got a major vote of confidence. Meta has signed a five-year agreement with Dutch-based Nebius to provide up to €27 billion in compute capacity, one of the largest infrastructure deals the social media giant has ever struck with an external provider.

What Happened

European neocloud Nebius has signed a $3 billion deal with Meta to provide the hyperscaler with AI infrastructure over a five-year period. More precisely, under the agreement, Nebius will provide $12 billion of dedicated capacity across multiple locations based on one of the first large-scale deployments of the NVIDIA Vera Rubin platform, with Meta committing to purchase additional available compute capacity up to a total of $15 billion over five years.

The company recently launched a 4,000-strong Nvidia B300 GPU cluster at Ark Data Centres’ Longcross Park campus in Surrey. The announcement follows a recent expansion of Nebius’s first data center in Mäntsälä, Finland, with the Nordic country emerging as a preferred destination for data center investment due to its low energy costs and cool climate.

Why This Matters for Europe

This deal represents a strategic shift in how major tech companies are approaching AI infrastructure in Europe. Nebius is GDPR-compliant with privacy-focused architecture and tenant-level isolation as standard. At a time when the EU AI Act is reshaping regulatory requirements for AI providers, having GDPR-compliant infrastructure from a European provider offers both compliance assurance and data sovereignty—concerns increasingly important for European enterprises and regulators.

The EU’s AI Act, which began enforcement in 2025, imposes requirements on providers of general-purpose AI models and the infrastructure that supports them, and as data center regulations tighten across Europe, the geographic distribution of Nebius’s infrastructure investments will have compliance implications.

Netherlands-based Nebius has emerged as a leading European player in the rapidly developing AI cloud computing space. Founded in 2022 after a restructuring of Russian company Yandex’s operations based outside of its home market, Nebius listed in New York in 2024.

What This Means for Builders

For developers and enterprises in Europe, this signals that major AI workloads no longer require exclusive reliance on US-based infrastructure. Meta cited “building a more resilient and flexible infrastructure” as the reason for the diversification strategy. This has immediate implications: European teams can now plan for GDPR-compliant training and inference at hyperscale, potentially reducing latency and compliance friction for regulated workloads.

For smaller European AI companies, Nebius’s growth creates a regional alternative to CoreWeave, Lambda, and other emerging infrastructure providers—one that understands EU regulatory requirements from the ground up.

Open Questions

The deployment timeline remains partially unclear. Nebius will deliver this capacity starting early 2027. How quickly capacity ramps and whether pricing becomes competitive with US alternatives will determine whether this becomes the default choice for European AI builders or remains a compliance-first option.

Also worth watching: whether other hyperscalers follow Meta’s lead in diversifying to European providers, or whether this remains a Meta-specific play.

Source: Data Center Dynamics