According to Crescendo AI, Meta has signed a five-year, $27 billion compute agreement with AI infrastructure provider Nebius. The deal includes $12 billion for dedicated infrastructure featuring one of the first large-scale deployments of NVIDIA's Vera Rubin GPU platform — the successor to Blackwell, offering approximately 3.3x better performance for large-scale AI training — plus $15 billion for additional capacity. Nebius operates across Europe and North America, and the agreement gives Meta GDPR-compliant compute for workloads requiring European data sovereignty. The infrastructure is intended to support training of Meta's next-generation Llama models and Orion multimodal systems. The deal reflects Meta's continued strategy of diversifying compute procurement across multiple providers rather than depending on a single cloud vendor.