IREN's latest results mark a concrete change in the economics surrounding Bitcoin mining. For the quarter ended June 30, the company reported $70.5 million in AI Cloud Services revenue and $66.7 million in Bitcoin Mining revenue. AI Cloud therefore became the larger of the two businesses for the quarter, even though mining remained dominant across IREN's full fiscal year.
The crossover matters because Bitcoin miners did not build only mining operations. They assembled power, land, grid interconnections, substations, cooling and data-center capacity. IREN's results show that some of those physical assets can serve another customer: artificial-intelligence computing.
What changed in one quarter?
For the quarter ended March 31, IREN reported $33.6 million in AI Cloud Services revenue and $111.2 million in Bitcoin Mining revenue. One quarter later, AI Cloud revenue had risen to $70.5 million while mining revenue had fallen to $66.7 million.
That is a current operating threshold, not a projection. It shows that infrastructure developed around Bitcoin mining can produce more quarterly revenue from a different computational workload.
Bitcoin remains central to the story. Mining helped finance and accelerate the development of the energized sites that AI companies now value. The crossover does not remove Bitcoin from IREN's history or business. It reveals a second possible economic life for infrastructure assembled around it.
Why is this not a simple AI victory story?
Across fiscal 2026, IREN reported $578.2 million in Bitcoin Mining revenue and $128.8 million in AI Cloud Services revenue. Mining remained substantially larger on a full-year basis.
The conversion also carried major costs. IREN reported a $684 million fourth-quarter net loss and $450.4 million in fourth-quarter non-cash impairments. The company said the impairments were primarily associated with decommissioning Bitcoin mining hardware as sites were converted to support AI Cloud growth.
The latest quarter therefore establishes a revenue crossover, not uncomplicated profitability. It does not prove that every Bitcoin miner can economically convert to AI or that AI demand is already weakening the Bitcoin network.
What is the breaking economic question?
IREN reported that recent three-year AI contracts exceeded $20 million of revenue per megawatt of IT capacity. That is a company-specific metric, not a universal price for power.
It still sharpens the question facing the sector: what happens when a megawatt controlled by a Bitcoin miner can be allocated to AI compute instead?
The answer will vary by site, power agreement, cooling system, network connection, financing structure and customer demand. IREN supplies one verified case. The wider industry outcome remains unresolved.
For now, the strongest conclusion is narrow and consequential. A Bitcoin mine can have a second customer, and IREN has shown that AI Cloud can overtake mining as a quarterly source of revenue without making Bitcoin irrelevant to the infrastructure story.