Image Credit: Shutterstock
Key Takeaways
- IREN’s co-founder says AI’s biggest bottleneck has shifted from chips to physical infrastructure like power, land, and cooling.
- The company has a $3.4 billion NVIDIA cloud contract and a $9.7 billion Microsoft deal anchoring its pivot from Bitcoin mining to AI infrastructure.
- IREN says it has secured roughly 5 gigawatts of grid-connected power globally and is targeting 150,000 GPUs by end of 2026.
IREN co-founder and co-CEO Daniel Roberts said the primary constraint on artificial intelligence is no longer semiconductor supply but physical infrastructure: power, land, cooling, and data center capacity. Roberts laid out the company’s strategy in an X post on Friday, framing IREN as a vertically integrated AI infrastructure platform spanning power generation through enterprise software.
Roberts Describes a Three-Layer Strategy
Roberts organized IREN’s business around three layers. The first is physical infrastructure, including power and data centers. The second is compute infrastructure, covering NVIDIA GPUs and servers. The third is enterprise software and operational tooling.
“AI demand grows exponentially. Infrastructure doesn’t,” Roberts wrote. “Layers 1 and 2 are where the overwhelming majority of IREN’s value is being created today. Layer 3 is where that advantage compounds further over time.”
The company, formerly known as Iris Energy, has expanded beyond its Bitcoin mining origins into AI cloud services across sites in Texas, British Columbia, Oklahoma, Spain, and Australia. Roberts said IREN has secured roughly 5 gigawatts of grid-connected power capacity globally. He described that figure as secured capacity, not a pipeline target.
A $3.4 Billion NVIDIA Contract and a $2.1 Billion Investment Right Anchor the Pivot
The infrastructure thesis is backed by a recently announced strategic partnership with NVIDIA. On May 7, the two companies said they would collaborate on deploying up to 5 gigawatts of NVIDIA DSX-aligned AI infrastructure across IREN’s global data center pipeline.
Separately, IREN signed a five-year, $3.4 billion AI cloud contract with NVIDIA for air-cooled Blackwell GPU deployments at its Childress, Texas campus. As part of the broader partnership, NVIDIA received a five-year right to purchase up to 30 million IREN shares at $70 per share, a potential $2.1 billion investment that vests only as GPU infrastructure is deployed across IREN campuses, with full vesting at 600,000 GPUs.
NVIDIA CEO Jensen Huang called AI factories “foundational infrastructure for the global economy” in the joint announcement.
IREN Is Also Building for Microsoft at Scale
The NVIDIA deal sits alongside a separate $9.7 billion contract with Microsoft at IREN’s Childress campus, where the company is advancing liquid-cooled data centers. IREN said it has increased annualized recurring revenue under contract to $3.1 billion and is targeting $3.7 billion by the end of calendar year 2026.
The company ordered more than 50,000 NVIDIA B300 GPUs earlier this year and expects to expand its total fleet to 150,000 GPUs by the end of 2026. IREN also acquired Spanish data center developer Ingenostrum, adding 490 megawatts of grid-connected power in Europe and bringing its total portfolio to the 5-gigawatt figure Roberts cited.
WhiteFiber Announces Separate AI Compute Deal in France
Separately, WhiteFiber announced a five-year AI compute agreement worth more than $160 million with an investment-grade technology customer in France. The deployment will use NVIDIA GPUs and expand WhiteFiber’s European footprint. WYFI shares rose 22% on Thursday and gained another 5% in Friday premarket trading.
WhiteFiber provides AI cloud and high-performance compute services using third-party data center infrastructure. IREN focuses on owning and operating the underlying physical infrastructure itself, a distinction Roberts emphasized in his post as the source of the company’s long-term moat.
