LIVE IREN Q3 Earnings | Opendoor and The Trade Desk
- The Microsoft AI Cloud contract is $9.7B total and is expected to deliver ~ $1.9B in annual revenue once four phases are live, with an estimated ~85% project EBITDA margin, supporting strong unlevered and levered returns.
- Management projects the GPU/cloud capex at $5.8B (GPUs and ancillaries); 20% customer prepayments fund ~one-third of initial capex and securitization of GPUs/contracted cash flows could raise ~ $2.5B.
- Q1 FY26 operating performance: total revenue $240M (up 28% QoQ, 355% YoY) and adjusted EBITDA $92M, though operating expenses rose due to higher depreciation and share-based payroll taxes.
- Data center supply-side posture: Childress engineered for single 100 MW superclusters and 130–200 kW per rack densities; Sweetwater planned up to 2 GW (2,000 MW) of build-out capacity secured from the grid.
- Financing posture: issued $1B zero-coupon convertible notes, secured $400M in GPU financing to date, and ended October with ~$1.8B cash; remaining capex to be funded via mix of cash, operating flows, GPU-backed lending, and equity/debt.
- Large Microsoft AI Cloud contract ($9.7B) unlocking ~$1.9B annual revenue and strong contracted cash flows.
- Very high projected project economics: ~85% project EBITDA margin and management-cited 35%+ equity IRRs.
- Scalable assets: Sweetwater (2 GW) and Childress (multi-horizon, up to 450 MW secured power) enable massive growth.
- Strong demand for high-density, liquid-cooled GPU capacity (NVIDIA GB300/NVL72) and pre-contracting from customers.
- Recent financings ( $1B convertible notes, $400M GPU financing ) and $1.8B cash position support near-term capex.
- Significant remaining funding need (~$1.4B) reliant on capital markets, securitizations, or equity issuance.
- Management left some analyst questions unanswered on the call, indicating disclosure and clarity limitations.
- Higher-spec design increases incremental capex and operating costs, compressing near-term margins.
- GPU refresh and auxiliary equipment may require future retrofits, creating reuse uncertainty and upgrade costs.
- Concentration risk: large exposure to a handful of hyperscalers could leave recontracting risk at term end.