The $9.7B deal highlights how AI demand is outpacing compute supply — and why Microsoft keeps signing billion‑dollar contracts.
As reported by Reuters, Microsoft, almost renowned now for making billion-dollar AI deals after billion-dollar AI deals, is back with another huge agreement. This time, it is partnering with IREN, with Microsoft prepaying a hefty sum upfront to secure access to IREN’s NVIDIA-powered servers as they come online.
The deal gives Microsoft more compute for AI workloads such as Copilot, OpenAI training, and Azure. It also lets it skip the slow, expensive process of building new data centres, which can take years and often get delayed by planning approvals and local power limits.
By working with IREN, Microsoft can scale far faster than if it tried to build everything itself. It also avoids buying GPUs that become outdated quickly, since Nvidia releases new hardware at a rapid pace. Renting cutting-edge chips through partners keeps Microsoft flexible and avoids being stuck with last-generation hardware.
Inside Microsoft’s $9.7B IREN contract
Microsoft HQ (Image credit: Windows Central)
The five-year agreement, costing $9.7 billion and including a 20% prepayment, will see Microsoft prepay IREN for cloud and GPU capacity. This locks in priority access to high-end NVIDIA hardware as it becomes available.
Microsoft will gain access to IREN’s new GPU clusters built around NVIDIA’s upcoming GB300-class chips, with immediate use of each server batch as soon as installations are completed. The GPUs will be deployed at IREN’s 750 MW Childress, Texas campus, with the new liquid-cooled data halls providing around 200 MW of critical IT load for Microsoft once everything is fully built out.
IREN is purchasing roughly $5.8 billion worth of Nvidia hardware and supporting infrastructure through Dell, which acts as the systems integrator. Dell supplies the full GPU servers and all surrounding equipment. This means everything arrives validated, certified, and ready for deployment, allowing Microsoft to slot the hardware straight into IREN’s racks without extended configuration or testing delays.
IREN also noted that it will fund the rollout using a mix of cash on hand, customer prepayments, operating cash flow, and additional financing. This helps keep the build moving at the pace Microsoft needs for its AI expansion plans.
Microsoft has already said that demand for Copilot and Azure AI is outpacing available compute, which is part of the reason we keep seeing so many billion-dollar AI agreements from the company. For some, it might feel exhausting to hear Microsoft announce yet another eye-watering deal, but it has quickly become the norm as the company goes all in on an AI future and works to stay ahead of rival cloud providers.
Who is IREN, and why is Microsoft working with it?
IREN is a rapidly expanding data center provider that specialises in housing large-scale GPU clusters. It already has the power capacity and infrastructure foundations in place, which means Microsoft can plug into new compute far sooner than if it tried to build a site on its own.
Why does Microsoft need so many GPUs right now?
AI features such as Copilot, Azure AI services, and OpenAI model development require a massive amount of computing power. Microsoft has said demand for these tools is rising faster than the hardware it has available, so it needs more GPUs to keep up.
What GPUs will Microsoft receive through the IREN partnership?
The contract gives Microsoft access to NVIDIA’s upcoming GB300-series GPUs. These units will be rolled out across IREN’s Texas facilities as new data halls come online over the next couple of years.
You can post now and register later.
If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.
3175x175(CURRENT).thumb.jpg.b05acc060982b36f5891ba728e6d953c.jpg)
Recommended Comments
There are no comments to display.
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.