Microsoft Locks In $9.7B AI Compute Deal with IREN to Ease GPU Shortages

• Microsoft signs a five-year, $9.7B deal with IREN for access to NVIDIA GB300-class GPU systems, easing AI compute bottlenecks.
• The agreement includes a 20% upfront prepayment and a $5.8B hardware procurement from Dell Technologies.
• IREN will host and operate GPU clusters at its 750 MW Childress, Texas campus, built for liquid-cooled, high-density systems.
• This deal helps Microsoft address AI GPU shortages expected to last through 2026 while expanding Copilot and Azure OpenAI services.

Microsoft has signed a five-year $9.7 billion agreement with IREN for access to NVIDIA GB300-class GPU systems. The arrangement includes an upfront 20% prepayment from Microsoft and a parallel $5.8 billion procurement by IREN from Dell Technologies for GPUs and supporting infrastructure.

Expanding compute without new data centers

Under the agreement, IREN will host and operate GPU clusters that Microsoft can use to handle its growing AI workloads in the next five years, giving Microsoft predictable access to high-performance compute resources without adding capital expenditures for new facilities.

Deployment will take place at IREN’s Childress, Texas campus, a 750 MW facility built with liquid-cooled data halls designed for dense GB300 systems. Initial phases are expected to provide around 200 MW of critical IT load, with phased expansions through 2026, depending on construction, power, and equipment milestones.

Easing the strain on AI capacity

Microsoft has cautioned that GPU shortages could continue into mid-2026, making hosted capacity a vital bridge for scaling Copilot and Azure OpenAI. For IREN, the partnership marks a major shift from its crypto-mining past toward AI and HPC cloud infrastructure, supported by secured power and advanced cooling technologies.

To fulfill the contract, IREN signed a $5.8 billion supply deal with Dell Technologies for NVIDIA GB300 systems and related infrastructure. This partnership positions Dell as a critical player in deploying liquid-cooled, high-density GPU clusters across North America.

Following the announcement, IREN’s shares jumped more than 20%, reflecting strong investor confidence in long-term AI infrastructure demand and the financial stability of having Microsoft as an anchor tenant.

Analysts are watching three key factors: the pace of the 2026 rollout at Childress, availability of NVIDIA GB300 systems amid global demand, and execution on power and cooling at scale in Texas.

IBS Insight

Mega deals like Microsoft–IREN shows the massive infrastructure demands behind AI growth from managing massive power loads, advanced cooling, and reliable interconnect performance at scale.

IBS Electronics can help you source the critical components you need from power semiconductors, passives, thermal management, and interconnect solutions while navigating lead times and cost pressures.

Talk to our engineers to optimize your BOMs for high-density, liquid-cooled systems that balance performance, reliability, and cost.