IREN IREN

Research status: Primary-source verified through FY2026 results on 27 August 2026. Current market reference: $44.68 close on 4 September 2026.

FY2026 results materially strengthened the demand and funding evidence while confirming that IREN remains a capital-structure-sensitive execution story. Management reported $4B of contracted ARR for 2026 capacity, with $1B of ARR operating today as of 26 August. The 2026 capacity is described as largely sold out, including successful delivery of Horizon 1 to Microsoft, and the customer base has broadened to hyperscalers, enterprises, AI developers and frontier labs.

The financing picture also improved: IREN disclosed approximately $14B of existing cash, committed GPU financing and customer prepayments, including $2.8B of GPU financings funding about 90% of associated GPU capex. This reduces the immediate common-equity-funding concern relative to the prior dossier, but does not eliminate it. ARR remains a non-GAAP operating metric and recognized revenue still depends on commissioning, testing, customer acceptance, utilization and contract terms.

At $44.68, the stock still offers substantial upside if contracted capacity converts on schedule and funding remains mostly non-dilutive, but the dispersion remains wide. Current conclusion: keep IREN at investigate/watch. The earnings gate improved the thesis but does not justify a buy plan without a fully reconciled per-share model for financing, dilution, depreciation/replacement capex and project returns.

IREN owns scarce power/data-center infrastructure and is successfully converting it into contracted AI-cloud capacity. FY2026 results strengthened the thesis by showing $4B of contracted ARR for 2026 capacity, $1B already operating, broader customer diversification and materially improved GPU financing. The central underwriting question remains whether this capacity build creates attractive per-share value after financing costs, dilution, depreciation and replacement capex. PowerFund will not equate ARR with GAAP revenue or equity value.

Conversion of $4B contracted ARR into recognized revenue; successful commissioning/acceptance of remaining 2026 capacity; sustained customer diversification; attractive utilization and project margins; additional customer prepayments/GPU financing; evidence that common-equity dilution remains modest; 2027 capacity contracting.

Extreme capital intensity; ARR-to-revenue conversion risk; commissioning/customer-acceptance delays; dilution and complex financing; GPU obsolescence/replacement capex; customer concentration; utilization/pricing compression; power/interconnection/cooling bottlenecks; accounting complexity during the transition from mining to AI Cloud.

Warning — freeze additions and investigate

• Material delay in Horizon / year-end 480MW delivery.
• Contracted ARR coverage falls below ~80% or major customer terms weaken.
• New deals require materially less customer prepayment and more parent-level equity funding.
• Fully diluted share count rises >15–20% without proportional increase in base-case per-share value.
• Utilization after customer acceptance trends below ~75–80%.
• Project-level returns are not clearly above cost of debt/equity.

Reduce

• Two consecutive material capacity/acceptance delays.
• Major contract cancellation or renegotiation reduces expected ARR by >15%.
• Financing requires repeated common-equity issuance at unattractive prices.
• AI Cloud EBITDA grows but free cash flow remains deeply negative after normalized maintenance/replacement capex.

Invalidate

> Invalidate if IREN proves able to build capacity but not create per-share economic value—specifically, if deployment requires persistent dilutive equity funding, utilization/pricing disappoint, or customer/financing terms drive project returns below a reasonable cost of capital.

Additional hard invalidators: loss of a major hyperscaler contract, inability to finance committed deployments, or structural deterioration in secured-power economics.

IREN competes with hyperscalers, NeoClouds and data-center/GPU infrastructure providers. Its differentiator is owned/controlled power and site development capability, combined with a vertically integrated cloud platform.

Advantages:

• large secured power pipeline;
• demonstrated ability to build dense compute sites;
• major strategic customers;
• increasing use of customer prepayments;
• ability to allocate capacity across customers/platform layers.

Disadvantages:

• enormous capital requirements;
• lower software moat than hyperscalers;
• hardware depreciation risk;
• financing complexity;
• potentially commoditizing GPU-hour economics.

The moat should be judged through time-to-power, deployment speed, cost/MW, uptime and contract economics, not by GPU count alone.

1. Reconcile FY2026 cash, restricted cash, debt/GPU financing, customer prepayments and fully diluted share count from the 10-K.
2. Build a site-by-site 2026/2027 commissioning schedule with MW, GPU type, customer, acceptance and billing dates.
3. Reconcile the $4B contracted ARR and $1B operating ARR to GAAP revenue quarterly; track timing slippage explicitly.
4. Build project-level economics for Microsoft and major frontier-lab contracts including GPU capex, prepayments, financing cost, depreciation/replacement capex, utilization and terminal value.
5. Quantify dilution from warrants/equity-linked instruments and test whether per-share base-case value rises faster than fully diluted share count.
6. Compare IREN expected return and financing quality with NBIS and other AI-cloud infrastructure alternatives before any buy plan.
7. Re-underwrite after the next quarterly update or earlier on a material financing, commissioning, or customer-contract change.