Another billion‑dollar bet: Microsoft spends $9.7B on NVIDIA GPUs to fuel Copilot and Azure AI

by

Microsoft, increasingly known for its large investments in artificial intelligence, has announced another major deal. This time, they’re partnering with IREN, paying a significant amount in advance to guarantee access to IREN’s servers, which are powered by NVIDIA technology, as soon as they become available.

This agreement provides Microsoft with increased computing power to support AI tasks like Copilot, training OpenAI models, and running Azure services. Importantly, it allows Microsoft to avoid the lengthy and costly process of constructing new data centers, which can face years of delays due to permitting and power availability.

Partnering with IREN allows Microsoft to grow much more quickly than if they were to develop everything on their own. It also saves them from the expense of buying GPUs that quickly become obsolete, as Nvidia frequently releases new hardware. By renting advanced chips through partners, Microsoft stays adaptable and avoids being stuck with older technology.

Inside Microsoft’s $9.7B IREN contract

Microsoft has made a $9.7 billion, five-year deal with IREN, including an upfront payment of 20%, to secure cloud and GPU resources. This agreement guarantees Microsoft early access to NVIDIA’s most powerful hardware as it’s released.

Microsoft will be among the first to use IREN’s new, powerful GPU servers, which are built with NVIDIA’s next-generation GB300 chips. Microsoft will start using these servers as soon as they’re installed at IREN’s data center in Childress, Texas. Once the entire facility is complete, it will provide approximately 200 megawatts of computing power for Microsoft, adding to the data center’s total capacity of 750 megawatts and utilizing advanced liquid cooling technology.

IREN is a rapidly expanding data center company specializing in powerful, AI-capable infrastructure. With locations across North America totaling 2,910 MW of power capacity, they’re well-positioned for growth. Because IREN already owns land, has secured power sources, and is actively building, Microsoft can expand its services with them quickly. Microsoft views IREN as a key partner due to their reliable power supply and comprehensive AI cloud platform.

IREN is buying about $5.8 billion of Nvidia equipment and related infrastructure through Dell, who is assembling the complete systems. Dell provides fully built and tested servers and all necessary supporting equipment. This ensures everything is ready to use immediately, allowing Microsoft to quickly install the hardware into IREN’s data centers without lengthy setup or testing.

As a tech follower, I’m really watching IREN’s progress, and it’s good to see they’ve got a solid funding plan in place. They’re using a combination of their existing funds, money already received from customers, and ongoing revenue, plus they’re open to additional financing if needed. This is crucial because it allows them to build out the infrastructure Microsoft requires to keep scaling its AI initiatives – basically, they can keep up with the speed Microsoft needs!

AI demand isn’t slowing

I’ve been watching Microsoft make some big moves to boost its AI capabilities. They recently signed a deal worth $17.4 billion with Nebius to get more GPU power. On top of that, they’ve partnered with CoreWeave and Lambda in a multi-billion dollar arrangement focused on renting GPUs and building dedicated AI data centers. It seems like these partnerships are all about helping Microsoft grow quickly enough to keep up with the soaring demand for AI.

Microsoft has reported that demand for its AI services, Copilot and Azure AI, is growing faster than their current computing capacity. This explains why the company is frequently announcing large, multi-billion dollar AI partnerships. While it may seem like a constant stream of big announcements, Microsoft is heavily investing in AI and aiming to lead the market in cloud technology.

Despite these collaborations, experts believe that GPUs will continue to be in short supply until at least 2026, because of high worldwide demand and the slow pace of improvements to power infrastructure.

FAQ

Who is IREN, and why is Microsoft working with it?

IREN is a fast-growing data center company that focuses on hosting powerful GPU clusters. Because they already have the necessary power and infrastructure, Microsoft can quickly access new computing resources – much faster than if they had to build a new facility from scratch.

Why does Microsoft need so many GPUs right now?

Tools powered by artificial intelligence, like Copilot and Microsoft’s Azure AI services, need a lot of computing strength. Microsoft has explained that demand for these tools is growing quickly, outpacing their current hardware supply, so they’re working to acquire more GPUs to meet the increasing need.

What GPUs will Microsoft receive through the IREN partnership?

Microsoft will gain access to NVIDIA’s next-generation GB300-series GPUs. These GPUs will be deployed in IREN’s data centers in Texas as new sections become operational over the next two years.

Read More

2025-11-11 18:40