IREN pushes $4 billion-plus annualized revenue target into 2027 after Microsoft changes

IREN will have to wait until 2027 to reach a contracted annualized revenue target of more than $4 billion. In its November 5 quarterly results, the AI cloud provider moved the milestone from year-end 2026 to January, possibly late February, citing a customer-requested design change and supplier timing. Co-CEO Daniel Roberts identified that customer as Microsoft.

A network change moves the revenue clock

Microsoft asked IREN to reconfigure and upgrade network infrastructure so that 200 megawatts of capacity could serve the latest generation of frontier AI models, Roberts told analysts, according to CNBC. IREN also cited supplier timing in explaining the revised schedule. The change delays when the company expects to put the contracted capacity into operation.

Some Microsoft infrastructure is already running. IREN said Horizon 1 entered service during the quarter, while Horizon 2 through 4 reached late-stage construction. Roberts said operating annualized revenue had doubled to $1 billion. Those operating deployments are distinct from the larger target whose timetable has now slipped.

The $4 billion-plus figure is an annualized run rate, or ARR—not revenue IREN expected to earn during 2026. IREN calculates it by multiplying hourly prices for contracted, commissioned graphics processing units, or GPUs, by a full year’s hours, then adding annualized storage and related revenue. It measures a projected yearly pace rather than sales already recorded.

The revised target is for that capacity to be operational by January 31, with a possible extension to the end of February. Revenue is expected to ramp in the months after each data center is delivered, subject to commissioning, testing and customer acceptance. IREN cautions that recognized revenue may be materially lower than its ARR metric.

AI revenue grows, with leases now in the mix

For the quarter ended September 30, AI Cloud revenue comprised $135 million in services and $65 million in leases. The preceding quarter had $71 million in services revenue and no lease revenue. Bitcoin mining contributed another $28 million, down from $67 million, as the company’s revenue mix shifted toward AI.

IREN remained loss-making despite the higher sales. Its quarterly net loss narrowed to $263 million from $684 million in the June quarter. Revenue backlog rose 22% to $20.3 billion, with about two-thirds tied to investment-grade counterparties. That backlog remains subject to delivery and service-availability requirements; it is not revenue already earned.

Quarterly revenue: June to September 2026

Total revenue

$137 million→$228 million

USD millions

Total quarterly revenue increased 66% sequentially.

AI Cloud revenue

$71 million→$200 million

USD millions

AI Cloud quarterly revenue increased 182% sequentially.

Fully funded capacity, but not the whole buildout

IREN says the delayed target is fully contracted and fully funded. Its definition of funded covers remaining data-center and GPU spending for the 0.3-gigawatt program supporting that target. The funding includes committed facilities and contracted customer prepayments not yet received, and remains subject to financing conditions. It does not mean every dollar is already available as cash.

The broader spending plan has a different funding position. IREN expects $15 billion in capital expenditures over the remainder of fiscal 2027, split between two categories:

• $9 billion from cash, committed GPU financing and customer prepayments.
• $6 billion from targeted GPU financing, prepayments, data-center financing and other corporate sources, subject to market conditions.

Customer advances are already part of IREN’s approach. After quarter-end, it signed a five-year contract with prepayments covering 90% of the associated estimated GPU capital costs. Those payments are contractually due before service delivery. IREN cautioned that prepayment terms vary and future contracts may not offer similar arrangements.