IREN's earnings release says its more than $4 billion revenue target is "fully contracted and fully funded." The same release moves the date for reaching that target from the end of 2026 to January 2027.
Both statements can hold. They answer different questions. Picture a project tracker with two columns, "money lined up" and "open for customers." The word "funded" sits in the first column and the new date sits in the second. A footnote tells you what filled the first one.
IREN sells access to AI computing power and also mines bitcoin. The release lists Microsoft and NVIDIA among its customers. The numbers below come from the copy of the release filed with the SEC.
What does "fully funded" mean in IREN's release?
IREN defines "fully funded" in footnote 5. Four listed sources cover the remaining building and chip costs of one program, if funding conditions are met.
Here is the footnote's own wording:
"'Fully funded' means these sources cover the remaining capex from September 30, 2026, including data centers and GPUs, for the 0.3GW (IT) program that supports the more than $4 billion ARR target, subject to satisfaction of the conditions to funding under committed financing facilities, including conditions to the release of escrowed proceeds and applicable availability periods."
"Capex" is spending on buildings and equipment. A GPU is a chip that runs AI work. "ARR" is annualized run-rate revenue, covered below. The "0.3GW (IT)" is the program's size, measured in the power the computing equipment uses.
The footnote counts four sources as of September 30. Each one carries a different catch.
| Source counted | Plain meaning | What the release says holds it back |
|---|---|---|
| Cash and cash equivalents | Money IREN has now | Nothing stated |
| Restricted cash | Cash held in escrow under IREN's loan and letter-of-credit arrangements | It "remains subject to contractual limits on withdrawal or use" |
| Undrawn committed GPU financing facilities | Loans lenders agreed to make that IREN hasn't borrowed yet | The footnote's "conditions to funding" and "availability periods" |
| Contracted customer prepayments not yet received | Money customers agreed to pay ahead of delivery | Not yet in IREN's hands |
Only the first row is money IREN can spend today with no strings. The other three are real commitments, and each has a step still to finish.
What is inside the $15 billion plan, and what is still a target?
IREN plans $15 billion of remaining spending in its 2027 fiscal year. Its release covers $9 billion with committed sources and calls the other $6 billion a target.
| Line | Amount | How the release describes it |
|---|---|---|
| Remaining planned capex in fiscal 2027, counted from September 30 | $15 billion | Footnote 11: the figure includes data centers and GPUs |
| Cash and committed GPU financing and prepayments | $9 billion | Uses the footnote 5 definition above |
| Target GPU financing, prepayments, data center financing and other corporate level sources | $6 billion | Described as "subject to market conditions, to drive future capacity targets" |
Our arithmetic: $9 billion is 60% of the plan and $6 billion is 40%.
The $6 billion is tied to future capacity targets. One is 0.8 gigawatts of data center capacity by the end of calendar 2027, which a second footnote says "may require additional capex in FY27." So the $6 billion looks like expansion money rather than a gap in the $4 billion program, though the release doesn't say so in those words.
The release does not say what the 0.3-gigawatt program costs, so the filing can't show whether $9 billion matches it. The $9 billion line carries the same footnote as the "fully funded" claim. That points to one pool of money, but it is our reading and not IREN's statement.
For another big number that depends on how it is defined, see our read of Anthropic's $518 billion plan.
Why did the date move, and to when?
IREN moved its target from the end of 2026 to January 31, 2027. It says the date may slip to the end of February after a customer-requested design change.
The release blames "a customer-requested change and supplier timing." Co-CEO Daniel Roberts put it this way: "after a customer-requested design change and supplier timing moved it from year-end; it may extend to the end of February."
Reuters, in a report posted on Euronext's site, said the customer was Microsoft and that it asked for network-design changes. Reuters also reported that IREN said the delay would not result in financial penalties.
The release itself lists a related risk. The timing of commissioning, testing and customer acceptance "may result in delay credits or other reductions in payments under our customer contracts." That reads as a general risk warning. We found nothing in the release saying any credits are owed.
Nexairi Dispatch
Get the next AI move before it turns into a workflow problem.
Join the free newsletter for concise AI news, practical checklists, and the decisions practitioners need to make next.
Free. No spam. You will also get the Nexairi Dispatch.
What is the difference between annualized revenue and revenue earned?
ARR multiplies contracted GPU prices by the 8,760 hours in a year. It is a run rate for chips switched on and under contract, not revenue IREN has booked.
Footnote 4 spells it out. The January 31 date is for being "operational." Revenue is "expected to ramp upon, and being subject to, commissioning, testing and customer acceptance in the months following each data center's delivery." The footnote adds that ARR "is an operating metric, not a U.S. GAAP measure" and that "recognized revenue may be materially lower."
Compare the sizes. IREN reported $200 million of AI cloud revenue last quarter ($135 million from services and $65 million from leases). Four times that is $800 million. Roberts says "operating ARR doubled to a billion." The target is more than $4 billion.
The comparison is rough, because ARR counts chips on a target date. It still shows how much has to switch on.
How to read the word "funded"
Money tied to conditions is not a warning sign by itself. Construction projects are often built on loans that release in stages and on customer prepayments. IREN's wording says as much once you read the footnote.
The narrower point is what "funded" means here. Here it means sources are identified and committed, with conditions still to meet. Cash in the bank is a different claim.
A fair alternative reading: a customer's design change can improve the finished product while it moves the date, and needing $6 billion for expansion doesn't show the existing program lacks backing. The next section lists what would settle it.
What would change each row, and where would you see it?
Each open item has a document that will move it. IREN's releases, its SEC filings and the IREN page on On the Record are the places to look.
| Open item | What would change it | Where it would show |
|---|---|---|
| Delivery date: January 31, 2027, may extend to the end of February | Commissioning and customer acceptance at each data center | IREN's next quarterly release |
| Funding conditions | Escrow released, facilities drawn, prepayments received | IREN releases and SEC filings |
| The $6 billion target | New financing announced, or not | IREN releases and SEC filings |
| Revenue against ARR | Recognized revenue after testing and acceptance | Reported revenue in later quarters |
What does the release show, and what is still unknown?
On paper, IREN has a 0.3-gigawatt program with committed sources behind it, a date that moved by one to two months and a wider plan that still needs $6 billion. Fermi's lease showed the same gap between a financing step and a closing.
What it leaves out is the program's cost, the funding conditions still open and whether any contract credits apply. Those are the answers to look for. The next checkpoint is January 31, 2027, IREN's own date for having the capacity operational.
Sources
Related Articles on Nexairi
Jim Smart is the founder and editor in chief of Nexairi. A Business Intelligence Developer with experience building data systems for Verizon, U.S. Army operations, and enterprise finance teams, Jim spent years turning complex data into decisions that executives could act on — dashboards, forecasting models, and automation pipelines across telecom and government contracting. He founded Nexairi to apply that same clarity to AI: making emerging technology understandable and actionable for the operators, accountants, and business owners who need it most. Jim holds GenAI certifications from the University of South Florida Bellini College of AI and completed Springboard's Data Science Career Track.
