TLDR
- Microsoft secured access to over 100,000 Nvidia GB300 chips through a $19.4 billion deal with neocloud provider Nebius Group NV.
- The computing power will support Microsoft’s internal teams building large language models and consumer AI assistants.
- Microsoft has committed more than $33 billion total to neocloud providers including Nebius, CoreWeave, Nscale, and Lambda.
- The strategy frees up Microsoft’s own data centers to provide AI services to paying customers while addressing capacity shortages.
- Microsoft’s first foundation AI models under Mustafa Suleyman were trained at a CoreWeave facility near Portland, Oregon.
Microsoft is going all-in on solving its AI computing problem. The tech company has locked down access to more than 100,000 of Nvidia’s latest GB300 chips through its partnership with Nebius Group NV.
The deal, worth up to $19.4 billion, was first announced on September 8. At the time, details were scarce.
Now the picture is clearer. The computing power will go to Microsoft’s internal teams working on large language models and consumer AI tools.
This is part of a bigger spending spree. Microsoft has committed over $33 billion to what the industry calls “neoclouds.”
These are small infrastructure companies that rent out AI-focused computing power. Beyond Nebius, Microsoft has deals with CoreWeave, Nscale, and Lambda.
Scott Guthrie leads Microsoft’s cloud operations. He described the company’s approach as “land-grab mode” in the AI space.
The goal is simple. Microsoft doesn’t want to run out of computing capacity.
The company faces a real shortage. Building new data centers takes time.
Neoclouds have already solved key problems like securing power supplies and chips. Renting from them speeds things up.
Freeing Up Resources for Customers
Here’s the strategy in plain terms. Microsoft uses neocloud facilities for internal work and OpenAI projects.
This frees up Microsoft’s own data centers. Those can then serve paying customers who need AI services.
The first foundation AI models built under consumer AI chief Mustafa Suleyman were trained at a CoreWeave data center near Portland, Oregon. This shows Microsoft is putting real work on these external systems.
The financial benefits matter too. Renting capacity lets Microsoft classify some expenses as operational rather than capital costs.
This affects cash flow, taxes, and how profits look to Wall Street. Analyst Mark Moerdler from Bernstein noted these accounting advantages.
The flexibility helps Microsoft pivot faster than if it only used its own facilities.
More Demand Than Competitors
Microsoft isn’t just training models at neocloud sites. Recent deals with Nscale in the UK and Norway will help deliver AI services in those regions.
Early in the AI boom, Microsoft rented capacity from Oracle to offer an AI version of Bing search.
Competitors like Amazon haven’t announced similar neocloud partnerships at Microsoft’s scale. Google is renting some CoreWeave capacity for its work with OpenAI, Reuters reported.
Guthrie explained why Microsoft needs more. The company has massive AI demand from products like GitHub Copilot and from OpenAI’s ChatGPT, which has hundreds of millions of users.
“Amazon doesn’t have anywhere near that scale. Nor does Google, in terms of total users with intensity, especially during business hours,” Guthrie said.
Microsoft isn’t slowing down its own infrastructure spending. On Tuesday, the company announced a second phase of development at a data center in Racine, Wisconsin.
The expansion will bring the site’s total utility power capacity to at least 900 megawatts. That’s almost as much power as a typical nuclear reactor produces.
Earlier this year, Microsoft paused or stepped away from multiple projects while evaluating its computing needs. Guthrie said the company will continue adjusting its infrastructure plans based on demand and regulations.
Stay Ahead of the Market with Benzinga Pro!
Want to trade like a pro? Benzinga Pro gives you the edge you need in today's fast-paced markets. Get real-time news, exclusive insights, and powerful tools trusted by professional traders:
- Breaking market-moving stories before they hit mainstream media
- Live audio squawk for hands-free market updates
- Advanced stock scanner to spot promising trades
- Expert trade ideas and on-demand support
