IREN’s AI Cloud Ramp Is Real, but the Next Handoff Slips

Results vs. Consensus

Q1 FY2027 scorecard

Revenue landed inside the range of published estimates. It was close to the $228.26m reference and above a separately quoted $217m benchmark. The stronger comparison was to our September thesis: we underestimated how quickly accepted GPU capacity could displace mining.

Year-over-year comparison

Sequential comparison

Revenue. AI Cloud added about $129m sequentially while mining gave up $39m, a replacement rate of about 3.3 times against 83% in June. Reported sales still fell 5% year over year, but the quarter showed that bringing leased and managed capacity online can more than offset the remaining mining teardown.

Margins. The $78m sequential gain in segment gross profit outweighed the $33m rise in cash operating expense. That expense, which includes R&D and excludes equity compensation, came in below our previous $125m to $135m cash SG&A expectation despite the broader cost scope. September’s leverage was real, but the next $50m to $60m of cost arrives before the next major revenue step.

EPS and cash quality. The sequential decline in GAAP net loss was largely explained by the $448m reduction in impairments. The prior-year comparison also includes a $665m unrealized gain on financial instruments, against a $28m loss this quarter. Operating cash flow was $529m, including a $645m deferred revenue increase. Excluding that movement as a cash-flow sensitivity, operating cash flow was negative $116m. Property and hardware purchases were $4.82bn; free cash flow on that capex definition was negative $4.30bn.

Segment Performance

Segment gross margins exclude depreciation and corporate expenses; individual revenue lines may not add to the rounded consolidated total.

AI Cloud: Customer acceptance starts to show in revenue

Services generated $135m against $71m in June, and leases added $65m, primarily after Microsoft accepted Horizon 1. Mirantis contributed after its August 3 close, but its results were immaterial to the consolidated statements. With about 88% of sales now in AI Cloud, the mix benefit from retiring miners is largely in the base. From here, higher earnings depend on adding billable capacity at attractive rates while controlling deployment and hardware costs.

“AI Cloud gross margin before depreciation and amortization was 86%, consistent with the prior quarter.” — Anthony Lewis, CFO

Assessment: We see a durable segment gross margin in the mid-80s as leases and managed services scale. IREN still has to recover the cost of the GPUs and facilities, so that margin supports the revenue ramp more clearly than it supports a high technology valuation.

Bitcoin Mining: The final dollars are lower margin

Mining sold output from 402 Bitcoin mined, against 2,039 a year earlier. Average operating hashrate fell to 8.8 EH/s from 45.3 EH/s. Lower coin output accounted for $116m of the $205m year-over-year revenue decline and lower realized average Bitcoin prices accounted for the other $89m. The remaining operation produced roughly $10m of segment gross profit as it kept incurring power and operating costs.

“Total revenue was $228 million, up 66% quarter-over-quarter, with Bitcoin mining revenue falling to $28 million as we continue to actively decommission Bitcoin mining hardware ahead of GPU installations. We aim to substantially complete this transition by the end of this quarter.” — Anthony Lewis, CFO

Assessment: We expect mining to contribute about $15m of December sales and little thereafter. A separate $21m mark on held-for-sale assets exceeded the $2m impairment charge, so a small impairment line does not mean the exit is costless. The October resolution of the contested ~$100m U.S. tariff removes one cash risk from the wind-down.

Key KPIs

Key Topics & Management Commentary

Overall Management Tone: Management was direct about missing the December milestone and more explicit than in August about the amount still to fund. It sounded confident on customer demand and more willing to defer projects rather than accept whatever financing terms the market offered.

1. The handoff is now a network and commissioning problem

Our September upgrade treated the timely Horizon 1 delivery as evidence that IREN could complete Horizons 2–4 by December. After the first handoff, Microsoft requested a network redesign so the 200 MW cluster could serve newer frontier models. Management also flagged outstanding supplier deliveries. Horizon 2 had its GPUs racked and was in commissioning, Horizon 3 had GPUs on site and its halls energized, and Horizon 4 was still awaiting GPU deliveries.

“If suppliers deliver, we're targeting January. If they're late, as suppliers have been across the industry this year, it may run to the end of February.” — Daniel Roberts, Co-Founder and Co-CEO

Assessment: IREN can accommodate a change request across the physical cluster because it controls the site. That does not insulate revenue from acceptance or supplier timing. On the roughly $3bn increment from the $1bn exit rate, a month is about $250m of run-rate-equivalent revenue before recognition effects. We take out roughly one month of that increment from our March-quarter case. The revised company milestone meets the economic downgrade condition we set, without establishing a contractual breach or the same delay at each hall.

2. The order book is deeper; availability still determines collections

The quarter-end $20.3bn backlog can be reconciled at reported precision to $6.347bn of service obligations and $13.957bn of remaining contracted leases. Of the service obligations, $1.529bn was expected within 12 months of September 30. For leases already reflected in the minimum-payment schedule, the FY27 remainder carried $271m of minimum payments. The schedule excludes availability-contingent revenue and does not reflect the full set of contracts for capacity still being delivered.

“Existing customers are renewing and expanding with us. New logos are proving demand, and the renewals are proving ongoing delivery.” — Daniel Roberts, Co-Founder and Co-CEO

Assessment: Fifteen customers and the frontier lab expansion support the September demand thesis. The first lease revenue also shows that some contract economics depend on service availability: $49m was fixed or in-substance fixed, and $16m was subject to availability conditions. We give the signed book more weight than new project announcements, but value it on billable months and margins. Three customers accounted for 70% of quarterly sales, so a broader roster has yet to reduce exposure to a large buyer’s timetable.

3. New contract economics meet the pricing test

Pricing on three-year contracts signed in the quarter was $19m to $23m per IT MW, with five-year deals at roughly $17m per MW. A post-quarter five-year deal cleared about $22m with a contractual prepayment covering roughly 90% of GPU capex, meeting the threshold we set in September. The separate short-duration contract around $40m per MW offers higher annualized pricing, but leaves earlier renewal risk and should not be applied to multiyear deals.

“On a five-year contract at $20 million per megawatt and with a prepayment covering 50% of GPU CapEx, we get a payback on the GPUs of 18 months, and if we include the cost of the data centers, the payback is around three years. These are illustrative, and deal economics and payback periods will ultimately vary by customer, hardware, and contract structure.” — Anthony Lewis, CFO

Assessment: The combination of price and prepayment matters more than the higher quoted rate. On a customer-funded deployment, less equity is tied up before the first invoice, but prepaid cash is collected against future service and leaves less to collect later. We use $15m per incremental revenue-producing MW in the FY28 base case to allow for mix, availability and discounts on large anchors. The recent deals provide upside to that assumption without repricing the existing book.

4. IREN cut the capital plan and showed what still needs funding

Roughly $5bn spent plus $15bn remaining implies roughly $20bn for FY27, down from $25bn to $30bn. Management's $9bn of cash and committed financing and prepayments covers about 60% of the remaining plan. The funded claim applies to the initial 0.3 GW, subject to escrow releases, availability periods and other draw conditions. New Childress air-cooled deployments are spread across calendar 2027 as financing becomes more selective.

“Our planned CapEx for the remainder of the fiscal year is approximately $15 billion. This is down from the prior estimate, reflecting the flexibility to pace data center and GPU CapEx. With cash and committed funding sources of $9 billion, that leaves $6 billion additional planned CapEx to be funded.” — Anthony Lewis, CFO

Of the $5.78bn total cash pool, $2.26bn was restricted. The $2.45bn Mackenzie facility bears a 9% fixed rate, amortizes over 30 months from each draw and has parent recourse. Prepayments plus facility capacity can cover more than the equipment cost, but that leaves short amortization against a business still ramping. Management identified $12bn to $15bn of potential financing before equity or converts, and no completed data center financing was disclosed.

Assessment: Lower spending reduces the amount IREN must raise immediately, at the cost of a slower start to FY28 sales. The first financing against stabilized Horizons would change the equity case by demonstrating how much of the $6bn can be funded without issuing shares. Total contractual commitments rose to $18.95bn, and some orders can be cancelled only on contractual terms. The ability to slow projects is useful, but it is not costless.

5. Dilution is already visible in the share count

Stock compensation reached $106m, up from $43m in June, even though cash operating expense grew by only $33m. Quarter-end ordinary shares including loan-funded shares were 406.9m against 380.7m in June. The October 30 count was 424.4m, following another 17.4m ATM shares sold after quarter-end for roughly $630m gross, about $36.26 per share on average.

“In addition to data center and GPU financing activities, you can expect other corporate-level initiatives over time, and we will continue to maintain an appropriate balance of debt and equity funding as the platform scales.” — Anthony Lewis, CFO

Assessment: The sale average does not establish that the prior below-market trigger fired; it does show investors own less of each future dollar of cash flow. At $21.64, a $1bn equity raise would require roughly 46m shares before fees. We include future equity in the base case and do not add the full contingent securities balance to the denominator while also keeping convertible debt in enterprise value.

6. Secured power remains the asset that buys time

IREN retains more than 5 GW of secured power. Sweetwater 1 is already grid-connected and the 2 GW hub has conditional ERCOT Batch Zero baseload inclusion. Liquid-cooled construction is underway at Sweetwater 1 and has begun at Mackenzie; Spain’s 300 MW Badajoz project obtained a civil works license. These are different stages of development, with different dates for revenue.

“Many of those are early stage, and not all of them will convert to grid connection agreements.” — Daniel Roberts, Co-Founder and Co-CEO, on the development pipeline

Assessment: Sweetwater is evidence that early grid access can preserve a future build in a tighter market. More than 35 GW of projects under development still lack the commitments needed to underwrite revenue, and there is no named Batch Zero resolution for Horizons 5–6. We leave that wider pipeline out of base value. The benefit today is the ability to prioritize sites where power, tenants and financing arrive together.

Guidance & Outlook

The company did not issue formal quarterly revenue or EPS ranges, or a specific operating ARR target beyond the initial milestone. The guide gives a useful bridge to December: ~$250m from the existing run rate, plus residual mining and other effects. We expect about $265m of revenue, up around 16% sequentially. At an 82% gross margin before depreciation and $173m of cash operating expense at the midpoint, December Adjusted EBITDA drops to roughly $44m before other items. That is still positive, but well below September’s $66m.

The March quarter now needs substantially more billable capacity partway through the quarter. January acceptance leaves roughly two months for most incremental contribution; February leaves closer to one. Management gave a clearer calendar range than in August. The risk remains asymmetric near term: accelerating a signed deployment by weeks helps, but a late acceptance against the same staff and hardware costs can erase a material part of quarterly profit.

Analyst Q&A Highlights

What the Microsoft change order does to handoffs

The question was whether the customer change and supplier delays were distinct. The response identified the post-handoff network redesign and remaining components, but did not allocate the delay between them.

Q: “Maybe parse out how much of the supplier delay related to the change orders, or if they're unrelated, and just any other clarity around some of that commentary we've been hearing in the marketplace.”
— Paul Golding, Macquarie

A: “As we announced back in August, Horizon 1 was handed over to Microsoft, and subsequent to that, they made a request to reconfigure and upgrade some of the network infrastructure. The purpose of that was really to enable the 200 MW of capacity to service these latest generation frontier models. […] As a result, there will be a small delay. We're targeting January, as we noted in the release. It does depend on supplier timing, and it's been a tough year for suppliers and reliability of delivery. There is every chance that moves to the end of February.”
— Daniel Roberts, Co-Founder and Co-CEO

Assessment: A better cluster can be in the customer’s interest and still defer billing. Management described the extra capex as immaterial. The agreement extends delivery dates without delay credits when a Microsoft-requested change causes the delay. The parties are still discussing the dates and how fees, credits and reimbursements will apply. We treat this mainly as deferred billing and retain downside if supplier timing slips again.

Customer credit and the role of prepayments

The question tested whether a large deposit can substitute for a stronger borrower when IREN finances a deployment. Management described balancing large investment-grade contracts against more heavily prepaid AI developer deals.

Q: “But in this environment where funding is a concern, is having more IG-backed customers an emphasis for you, or do you not view it that differently if you're signing deals with customers that are prepaying for the entirety of the CapEx in some cases?”
— Brett Knoblauch, Cantor Fitzgerald

A: “So, we've got a $20 billion backlog of customer contracts at the moment, and it's been very deliberately constructed, where two-thirds of that is investment grade. So we're talking the likes of Microsoft, NVIDIA, $4 trillion credits standing behind that. Then, yes, there is a portion that is a lower credit, but they tend to go hand in hand with higher prepayments, as you mentioned, and are central to our broader strategic objective of servicing the broader ecosystem.”
— Daniel Roberts, Co-Founder and Co-CEO

Assessment: A deposit reduces IREN’s exposure before service begins. It cannot replace the value of a creditworthy counterparty for debt serviced from future rents, particularly if the facility has parent recourse. The larger prepayments improve project economics without removing the concentration that made the original funding package possible.

Why management is slowing capital expenditure

Asked why the plan was falling, management pointed to funding availability and cost. That explanation fits the decision to spread Childress commissioning through calendar 2027 even as signing prices improved.

Q: “Just as we think about that flexibility around the CapEx, what is kind of driving that? If you could provide some more color around why that looks like it is ticking down for this year.”
— Greg Lewis, BTIG

A: “We have deliberately flexed it down, and it is not because we want to, and it is not because we need to because of a lack of growth opportunities. The demand is there. The customers are there. The site, the 30,000 acres, the 5 GW of secured power, all the ingredients are there. It is purely and simply financial markets and the current state. Capital needs to be respected in the current environment, and we are not going to grow at all costs. We are going to find that middle ground based on the availability of funding, the cost of funding, to throttle that growth curve.”
— Daniel Roberts, Co-Founder and Co-CEO

Assessment: Financing conditions now determine which signed or signable opportunities IREN can pursue quickly. The cut can protect per-share value while still lowering near-term aggregate earnings; we model the slower Childress ramp rather than assuming the unspent dollars disappear from the cost of reaching 0.8 GW.

How much of the next 500 MW is contracted

The question sought the contracted share of the expansion after the new deals. The answer named customers and deployment phases without giving the share.

Q: “Just curious, what portion of the 500 MW of planned incremental data center capacity for FY 2027 is now contracted for after the recent deals you signed here?”
— Mike Colonnese, H.C. Wainwright

A: “In terms of the 0.5 GW, we have not guided specifically on that, the exact composition. But within that, it is clear that NVIDIA is a proportion of that capacity. It is clear that another good credit with one of these leading AI frontier labs is another credit behind that. So we have got the air-cooled at Childress that is rolling out in 2027, and then we have got VRs arriving later in the year.”
— Daniel Roberts, Co-Founder and Co-CEO

Assessment: The question used a fiscal-year premise; IREN's 0.8 GW total is an end-calendar-2027 capacity goal. The named demand supports a partial step-up, but without the contracted share, procurement dates or acceptance schedule we underwrite only 100 incremental revenue-producing MW on a full-year-equivalent basis in FY28.

Sweetwater’s grid status and permits

The question asked whether Sweetwater still needed final interconnection approval in December. Management distinguished Sweetwater 1’s connection from the broader ERCOT process.

Q: “Curious if you could remind us at Sweetwater, I know you have conditional baseload approval. Is your expectation still that you get final interconnection approval in December? Are there any other outstanding permits or just anything that we should think about in terms of bringing Sweetwater online?”
— Jonathan Petersen, Jefferies

A: “Just to be clear, at Sweetwater one, which is our 1.4 GW project, that is already connected to the grid. I am not quite sure what the December date was that you were referring to, but that project is already grid-connected. As we announced a couple of months ago, both Sweetwater one and Sweetwater two were included in the baseload of the batching process. That means all of the capacity, the 2 GW across those sites, is secured. […] I would also note for both Sweetwater one and Sweetwater two, those projects are already well under construction, as Dan pointed out during the presentation. You can see very clearly the status of those, but there is a significant amount of construction work on site already underway, and we obtain the relevant permits throughout the construction process.”
— Kent Draper, Chief Commercial Officer

Assessment: The physical connection removes one important gate for Sweetwater 1, while baseload inclusion protects queue treatment for the wider hub. Construction, permits and commercial commitments still determine when the next stages can earn revenue. Secured grid access therefore matters to relative delivery risk even before it earns a return.

Whether hardware inflation takes back the pricing gains

The question was whether rising GPU costs could erode the short payback periods. The chief commercial officer described pass-through in market pricing; the co-CEO added that IREN has deliberately left some orders open.

Q: “I guess, how do you expect those costs of purchasing GPUs to trend in the coming quarters? I guess in other words, what I'm trying to get at is, will the cost to purchase GPUs catch up to where you're leasing them at today, or do you expect these very short payback periods to persist for a while?”
— Jonathan Petersen, Jefferies

A: “I think it's well publicized that there continue to be supply challenges throughout the industry, in particular SSDs and RAM. There's various forms of memory have been constrained for quite a while. There are other components that continue to be constrained. In that environment, we would expect to continue to see GPU pricing increase over time. Obviously, that is driven by the end-user customer demand. Again, it's always interesting to see these dynamics where ultimately, the customer demand is just so strong that that is what is causing price increases in terms of hardware. What we see on the pricing side with customers is that any price increases flow through immediately to hourly pricing under contracts. The payback periods continue to at least remain where they are, and we've actually seen a shrinking of some of the payback periods over the past six months as that market demand continues to come through.”
— Kent Draper, Chief Commercial Officer

Assessment: Higher rental prices can offset higher equipment costs on new contracts, especially where deposits are large. IREN's risk disclosures qualify the claim for existing contracts, where input increases cannot always be passed through. Leaving some orders open protects liquidity if demand softens, while creating price and schedule risk if supply tightens.

What They’re NOT Saying

1. What a delayed handoff costs beyond foregone billing. Management characterized the network upgrade capex as immaterial but gave no schedule by tranche or amounts for the fees, credits and reimbursements still under discussion. A one-month slip is easier to model than a change to contracted economics.
2. Which non-equity funds arrive first, and on what terms. Indicative lending capacity does not specify advance rates, guarantees or when cash can be drawn. IREN did not reconcile the prior ~$8bn GPU funding and prepayment goal against the rephased capital plan.
3. What share of the next phase can earn through its hardware cost. Disclosed deal examples give pricing and prepayment but no capacity or segment earnings after depreciation. A handful of well-funded contracts cannot anchor margins across the next 500 MW.
4. How much more equity is required. The call outlined options but did not size the likely equity component, or the ongoing equity compensation cost after the founder and Mirantis awards. This matters more to per-share value as the share price falls.

Market Reaction

• Pre-print setup: IREN closed November 5 at $18.03, down 52.3% year to date against a 7.4% gain for the S&P 500. It had fallen 56.3% over 30 days and 76.4% over 12 months. The pre-print 52-week closing range was $18.03 to $76.41.
• After hours: Shares traded about 10% to 11% higher late in the November 5 extended session after the call.
• Reaction session: IREN opened November 6 at $21.15, up 17.3%, traded between $20.18 and $22.46, and closed at $21.64, up $3.61 or 20.0%. Volume was 112.9m shares versus a 51.2m 30-day average, or 2.2 times normal.
• Benchmark: The S&P 500 rose 0.3% in the reaction session, with AI infrastructure shares also recovering.

The rally followed a steep selloff as investors questioned financing for AI infrastructure and the path to new power approvals. The first lease revenue and evidence of higher contract prices gave investors reasons to reassess the worst outcomes, while the wider AI rebound helped. We read the move as relief that demand and initial delivery held up. A strong close does not establish that the delayed capacity or funding has cleared.

Street Perspective

Debate: Did the delivery outlook get worse, or just move a month?

Bull view: A customer requested the upgrade to serve newer models, the buildings are nearly finished and Horizon 1 is already earning revenue. A later commissioning date does not undo the contracted demand.

Bear view: Handoffs have slipped despite late-stage construction. Some bullish desks cut targets while keeping positive ratings because both acceptance and subsequent expansion revenue are coming later.

Our take: The economic damage is currently a partial-quarter loss of billing, not a loss of the contract. The cash cost is larger if slipped dates also defer access to financing. We value the March quarter below its full run rate and reserve the bearish case for a slip beyond the revised window.

Debate: Can owned data centers fund the next build?

Bull view: Owning the sites leaves collateral to borrow against and eliminates a landlord commitment that would require IREN to keep building into a weak funding market. Prepayments let some deployments fund most of their equipment.

Bear view: Lending against an owned site still requires stable tenant cash flows and acceptable terms. A smaller spend plan coincides with higher commitments, less free cash and more shares outstanding.

Our take: We expect IREN to use its collateral, but secured debt shifts cash flows to lenders before it helps common shareholders. A signed facility at a size that reduces likely equity issuance would do more for fair value than a new estimate of total borrowing capacity.

Debate: Is the latest pricing repeatable?

Bull view: A five-year deal above $20m per MW and a multiyear repricing of older H100s suggest tight capacity extends beyond the newest chips and short spot deals.

Bear view: Short-duration peaks can fall before IREN commissions the hardware, and legacy anchor agreements will not reprice with spot. Higher hardware costs can also absorb part of the benefit.

Our take: The five-year contract and multiyear H100 repricing are better underwriting evidence than the $40m short deal. Our upside case comes from putting more megawatts to work. Sustained higher contract prices would add to that upside, but one deal is insufficient to raise our base price assumption.

Model Update & Valuation Framework

Our estimates: Higher September revenue, fewer billable months in March

September sales were $63m ahead of the midpoint of our prior range. We had no published FY27 point estimate, but holding the other quarters fixed, a December handoff with the incremental capacity billable for the whole March quarter would yield about $1.0bn of Q3 revenue and $2.54bn for FY27. We now model $740m and $2.28bn respectively. End-February instead of end-January would put roughly another $250m of March-quarter sales and $210m of Adjusted EBITDA at risk before offsets. June assumes a full quarter from the initial program and a small amount from the next phase.

We estimate about $1.08bn of FY27 depreciation and amortization as installed GPUs and data centers enter service, $430m of stock compensation and $200m of net interest. On $1.22bn of Adjusted EBITDA, those costs imply a pre-tax loss of about $490m before marks and other reconciling items. Adjusted EBITDA improves well before accounting earnings; financing and hardware replacement costs still have to be paid.

For FY28 (ending June 2028), we underwrite about $4bn of recognized sales from the initial program once in service, plus $1.5bn from the equivalent of 100 additional revenue-producing IT MW for a full year at $15m per MW. That recognizes only a portion of the 500 MW expansion because the end-calendar-2027 construction target does not specify how much will be accepted or used for all of FY28. At an 84% segment gross margin and $1.54bn of cash operating expense as more sites ramp, $5.5bn of revenue produces $3.08bn of Adjusted EBITDA, or 56%. We estimate $2.15bn of depreciation and amortization: about $9bn of average in-service hardware at five years, $6bn of in-service buildings at 20 years, and $50m from other assets. These are modeled in-service balances, excluding land and construction not ready for use, and below management’s total asset target. Allowing another $480m of stock compensation and $650m of net interest leaves roughly a $200m pre-tax loss before marks and other items. The large difference is why we make a separate debt and dilution bridge.

12-month value at the November 6 close

We value FY28 operating earnings from the capacity we expect to be in service by then. The prior recap had no price target and compared enterprise value with the gross contracted book. That screen now misses a material swing in borrowing and share count. Here enterprise value is Adjusted EBITDA times the multiple. We subtract projected net debt including finance leases and the present value of services already paid for, then divide by projected shares. Our base multiple of 7.75 times Adjusted EBITDA corresponds to a roughly 4% operating earnings yield after depreciation and before equity compensation and tax. That is already a generous valuation for short-lived hardware. It assigns value to the longer-lived power and data-center assets and the earnings growth that can follow the contracted initial program. The lower downside multiple reflects less productive equipment and a higher cost of capital; the upside multiple requires the next phases to contract and finance on better terms.

Base capital bridge. September debt at principal including finance leases less unrestricted cash was $6.83bn. We add $16bn of capex through the horizon, including the $15bn remaining FY27 plan and $1bn early in FY28, and deduct $2.26bn of restricted cash assumed used for eligible investment and the $630m of disclosed post-quarter gross ATM proceeds. We assume $3.5bn of additional customer deposits, $1.1bn of equity and just $400m of operating cash after excluding new deposits. That cash figure starts from about $2.18bn of Adjusted EBITDA from October 2026 through the horizon, less roughly $400m of net interest and $1.35bn of prepaid revenue with no new collection, before small other cash movements. New debt funds the residual, leaving $14.94bn of net debt. Each additional $1bn reduces value by about $2.11 per share.

Share count and converts. We project 475m base shares: 424.4m in October plus 44m from $1.1bn of equity at $25 and about 6.6m for awards, including 4.55m from the first scheduled annual vesting of the co-CEO grants. We use roughly 7m to 8m total award shares across the cases. The main scenarios carry convertible debt at par and exclude the associated shares. The $445m of 2029 and 2030 notes met the share-price conversion trigger at September 30. If fully converted to shares at the initial terms, they would add roughly 30m shares and remove $445m of debt; upside value falls to about $32.21 before capped-call protection and other adjustments.

Prepayments reduce borrowing, then reduce collections. The enterprise multiple values revenue as IREN delivers the service. To avoid also counting the cash already received for those future services, we start with the $2.10bn September deferred lease revenue balance as a proxy for cash-backed deferrals, add each case’s modeled deposits and deduct the prepaid revenue we expect IREN to recognize through November 2027: $1.15bn downside, $1.35bn base and $1.60bn upside. The base therefore leaves $4.25bn still to earn, reflecting staged commencements and multiyear terms. Discounting the remaining cash shortfall by 15% approximates two years at an 8.5% financing rate. This is an equity valuation adjustment for future revenue without a matching cash receipt, separate from accounting debt and the cost of service already in earnings.

Downside and upside. The downside assumes a later ramp, only $350m of incremental sales over the initial program at the same pricing equivalent, weaker absorption of fixed costs, a further pace cut and greater equity issuance at lower prices. We assume about $12bn of capex from September through the valuation date in the downside, $2.85bn of new customer prepayments, $1.75bn of additional equity at $24 per share, and roughly break-even operating cash after interest and recognizing prepaid revenue without a matching receipt. Net debt is lower because less equipment is purchased and more of the residual is raised as equity, yet debt and prepaid services together absorb nearly all the modeled enterprise value. The upside assumes about 170 additional full-year-equivalent revenue MW at $15m per MW, plus modest ancillary growth. We model roughly $16bn of capex, offset by $4.65bn of new customer deposits, $850m of new equity at $25 and $650m of operating cash after interest and recognizing prepaid revenue without a matching receipt. The same starting debt, eligible restricted cash and already disclosed ATM proceeds apply to both cases. We give the base 65% weight because the initial 0.3 GW is contracted and funded subject to draw conditions. We give downside 20% and upside 15%: a further delay or tighter credit can hurt on its own, while the upside requires commissioning, new contracts and attractive funding to line up. With no dividend, that gives about $12.54 of weighted value, or a 42% loss against our 6% to 8% S&P 500 return assumption. Even halving the prepaid-service deduction leaves about 24% downside. Moving five points of weight from the base case to upside leaves about 37% downside with the full deduction.

Thesis Scorecard Post-Earnings

Overall: The conversion thesis has more evidence behind it than in September. The calendar and funding assumptions that supported an Outperform now carry more risk. The case also depends on cash collection after acceptance, because the customer deposits used to fund construction reduce what IREN can collect as it reports the revenue. Management met the pricing test and kept the mining exit on course, but the revised January or February milestone breaks the original December operating commitment.

Action: Underperform. The December slip warranted at least the downgrade we laid out in September. The additional step reflects the valuation and cash burden, independent of whether management meets its previous $8bn funding objective on a comparable basis. We would revisit the rating after acceptance within the revised window, closed financing that materially reduces the need for equity, and evidence that collections after deposits can support debt service and replacement spending. Another slip into the June quarter or repeated low-price equity issuance makes the downside case more likely.