Microsoft is intensifying its race in artificial intelligence with a colossal US$60 billion investment into “neocloud” data center providers, third-party operators that build and run specialized computing facilities optimized for AI workloads.
The multi-year commitments mark one of the largest infrastructure expansions in the company’s history, underscoring its ambition to maintain dominance in the fast-evolving AI compute market.
The tech giant’s largest pledge, around US$23 billion, will go to Nscale, a UK-based firm providing Microsoft access to roughly 200,000 Nvidia GB300 chips across key sites in the UK, Norway, Portugal, and Texas. Additional contracts include up to US$19.4 billion with Nebius, over US$10 billion with CoreWeave, US$9.7 billion with Australian firm Iren, and more than US$2 billion with Lambda, securing long-term access to critical GPU capacity.
These contracts, many spanning five years, align with Microsoft’s hybrid approach of blending owned and third-party data centers to balance flexibility and cost-efficiency.
The move is part of a broader effort to meet surging demand for AI compute power, a trend accelerated by the growth of tools like Copilot, ChatGPT, and enterprise-grade AI models integrated into Microsoft Azure.
In its most recent quarter, Microsoft reported spending nearly US$35 billion on infrastructure, primarily directed toward data center leases, GPUs, and networking equipment. The fresh $60 billion commitment signals not only continued expansion but also a strategic diversification of supply chains in an increasingly constrained global chip market.
Nscale’s projects alone promise major scale-ups, with its Texas campus expected to go live in 2026 and capable of expanding to 1.2 gigawatts (GW). Similarly, Iren’s Childress site in Australia plans to add liquid-cooled phases totaling 200 MW through 2026, while Nebius’s 300 MW New Jersey facility is scheduled for completion in summer 2025.
While Microsoft’s neocloud push is aggressive, analysts warn that delivery timelines remain uncertain. Several of the projects are still in early construction phases, and their success depends heavily on the availability of power, cooling systems, and Nvidia’s next-generation GPU supply chain.
Contracts may also involve take-or-pay agreements, obligating Microsoft to pay for reserved capacity whether or not it’s used, or usage-based models, where payment depends on actual consumption. These terms have not been made public, leaving open questions about financial exposure if construction delays or hardware shortages arise.
Notably, Iren’s partnership with Dell to purchase $5.8 billion worth of GPUs and equipment introduces another variable in the supply chain. Delays tied to grid availability or construction could push some deployments beyond their scheduled timelines, potentially affecting Microsoft’s near-term compute goals.
Microsoft’s investments reflect a broader industry trend where cloud and AI providers are racing to secure compute dominance.
With rivals such as Google, Amazon, and Meta making parallel multibillion-dollar expansions, the focus has shifted from software development to hardware scalability and data center geography.
By anchoring its strategy around neocloud operators, Microsoft gains a geographically diverse network of AI-ready infrastructure, essential for latency reduction, compliance with local data laws, and resilience against global chip shortages.
The post Microsoft Deepens AI Investments with $60B Multi-Year Neocloud Commitments appeared first on CoinCentral.


