• Meta will purchase at least $12B in AI infrastructure capacity from Nebius starting in 2027.
  • The agreement could reach $27B over five years as companies race to secure computing power.

What Happened

Meta has signed a major infrastructure agreement with AI cloud provider Nebius to secure computing capacity for artificial Intelligence workloads.

According to the report, the deal centres on Nebius supplying dedicated AI computing resources to Meta as the company expands its AI development programmes.

Under the agreement, Nebius will deliver around $12B worth of dedicated AI computing capacity across multiple locations starting in early 2027. Meta has also committed to purchase up to $15B in additional computing capacity if it is not sold to other customers, bringing the potential value of the contract to roughly $27B over five years.

The infrastructure will rely on high-performance computing clusters built with advanced processors from NVIDIA. These systems are designed to train and run large AI models that require enormous amounts of processing power and energy.

Nebius, headquartered in Amsterdam, focussedes on providing specialized AI cloud infrastructure rather than general-purpose cloud services. The company has positioned itself as part of a growing group of “neocloud” providers that target AI developers and large technology firms.

The agreement builds on previous collabouration between the two companies. Nebius had earlier signed a $3B infrastructure contract with Meta in 2025 and other large deals with major technology companies seeking AI computing resources.

Also Read: https://btw.media/en/allit-infrastructure/Intel-forecast-shortfall-highlights-struggle-to-meet-ai-data-centre-demand/

Why It’s Important

The deal illustrates the rapid escalation of competition for AI computing capacity. Training large language models and other advanced AI systems requires enormous processing resources and specialized hardware.

Major technology companies increasingly rely on external infrastructure providers to supplement their own data centre capacity. Locking in long-term supply agreements helps companies avoid shortages of GPUs, electricity, or data centre space.

For Nebius, the partnership could accelerate the growth of its AI cloud business. Large multi-year contracts provide predictable revenue streams that can justify massive investments in new data centres and computing clusters.

Yet the surge in AI infrastructure spending raises broader questions. Building large computing facilities requires vast amounts of electricity, specialized chips, and capital investment. If demand slows or AI development shifts direction, some of these assets could become expensive to maintain.

The agreement therefore reflects both opportunity and risk. While companies like Meta push forward with large-scale AI development, the infrastructure race may reshape the cloud industry and determine which providers emerge as long-term leaders in the AI economy.

Also Read: https://btw.media/en/newstop-tech-stories-today-october-22-2024/