IREN — Earnings Review Q2 FY26 (Atlas)

Date: 2026-02-23 Quarter: Q2 FY26 (Oct–Dec 2025) Market cap: 12.1B|EV: 13.0B | EV/TTM Rev: 17.2x | EV/Contracted ARR: 5.6x Revenue: $184.7M (+59% YoY, -23% QoQ) | Adj EBITDA: $75.3M (41%)

Verdict

IREN is in a planned transition valley. Bitcoin mining revenue collapsed 28% QoQ (-65.5M)whileAIcloudgrew13710M) — the gap does not close until Microsoft revenues commence April 2026. The Q2 FY26 double miss ($184.7M vs $224M consensus; -17.7%) is a timing artifact: capacity is migrating from BTC to AI cloud faster than contracted AI cloud revenue is hitting the P&L. The core data point is $2.3B in contracted ARR vs $757M TTM reported revenue — a 3x gap that begins unwinding in Q3 FY26. Capital markets execution was the actual beat: $3.6B GPU financing secured vs $2.5B target, $9.2B total funding in hand, $2.8B cash. Conviction 3/5. The bull case is concrete (contracted, not forecast) but binary on Sweetwater 1 energizing on schedule in April.

Qualification Gate

Dilution gate failure is structural: IREN raised $9.2B to build infrastructure against $9.7B contracted revenue. GPU financing at <6% plus $1.9B Microsoft prepayment cover ~95% of GPU capex, limiting marginal dilution going forward. The worst of the share issuance is likely behind it. Monitor quarterly.

Six-Factor Score

The Numbers

Revenue — 10 Quarters

| | Q124 | Q224 | Q324 | Q424 | Q125 | Q225 | Q325 | Q425 | Q126 | Q226 | | | Sep-23 | Dec-23 | Mar-24 | Jun-24 | Sep-24 | Dec-24 | Mar-25 | Jun-25 | Sep-25 | Dec-25 | |---|---|---|---|---|---|---|---|---|---|---| | Revenue ($M) | 34.4 | 42.6 | 54.4 | 57.0 | 52.8 | 116.1 | 144.8 | 187.3 | 240.3 | 184.7 | | YoY % | — | — | — | — | +54% | +173% | +166% | +229% | +355% | +59% | | QoQ % | — | +24% | +28% | +5% | -7% | +120% | +25% | +29% | +28% | -23% | | Gross Margin % | n/a | n/a | n/a | n/a | 39.6% | 72.2% | n/a | 71.8% | 66.4% | 64.4% | | Op Margin % (GAAP) | n/a | n/a | n/a | n/a | -89% | +15% | n/a | +11% | -32% | -63% | | Adj EBITDA % | n/a | n/a | n/a | n/a | 4.7% | 54% | 57% | 65% | 38% | 41% | | Net Income ($M) | -5.3 | -5.2 | +8.6 | -27.0 | -51.7 | -21.9 | -16.3 | +176.9 | +384.6 | -155.4 | | EPS GAAP | n/a | n/a | n/a | n/a | -0.27|−0.10 | n/a | n/a | +1.08|−0.52 | | EPS Non-GAAP | — | — | — | — | — | — | — | — | — | -0.03||FCF(M) | n/a | n/a | n/a | n/a | -387 | n/a | n/a | n/a | -138 | -648 |

GAAP P&L distorted by unrealized gains/losses on financial instruments (Q126 GAAP EBITDA +662.7M; Q226−243.9M), debt conversion inducement (-$111.8M Q226), and D&A ($99.2M Q226, +175% YoY). Adj EBITDA is the cleaner operating signal.

10 quarters of data available. 12-quarter requirement not met — structural post-IPO constraint.

D&A and SBC (Selected Quarters)

| | Q225 | Q325 | Q425 | Q126 | Q226 | | | Dec-24 | Mar-25 | Jun-25 | Sep-25 | Dec-25 | |---|---|---|---|---|---| | D&A (M)|36.1|47.3|63.8|85.2|99.2||SBC(M) | 7.9 | 7.8 | 18.7 | 72.4 | 58.2 |

D&A up +175% YoY as asset base expands faster than revenue recognition. SBC spike in Q126 driven by performance RSU thresholds; $22.3M of Q226 SBC is accelerated amortization on performance instruments.

Non-Cash Items — Q226 GAAP Impact

Revenue Segment Mix

Prior Beliefs / Updated Beliefs

Prior beliefs entering Q2 FY26: BTC mining to hold ~$220M as AI cloud contributes 12 − 15Mofearlyramp.AdjEBITDA 90M at ~38% margin. GPU financing target ($2.5B) to be confirmed. Capital markets solid, no upside surprise expected.

Delta: Revenue miss is real but misattributed. The market labeled this a "double miss" — it is more accurately a timing miss in a deliberate transition. Bitcoin mining fell harder and faster than modeled as IREN proactively migrated capacity, while AI cloud (+137% QoQ) is inflecting. Capital markets execution materially over-delivered: $9.2B secured in a single fiscal year. The Q3 FY26 setup is better than I entered Q2 expecting — more capital, faster GPU deployment optionality, Sweetwater on schedule. The bearish reaction to Q2 FY26 results is a better entry window for the thesis, not a thesis break.

Leading Indicators

The largest contracted-to-reported divergence in this portfolio. $2.3B contracted ARR vs $69M annualized AI cloud run rate — 33x gap. Begins closing in April 2026.

Bullish divergence: sustained and widening. Contract signed → GPU financing secured → Microsoft prepayment received → revenue not yet reporting. The lead time between contract and revenue recognition is the buy window. This is not a leading indicator trend — it is contractually guaranteed revenue that has not yet hit the P&L.

Watch in Q3 FY26: Sweetwater 1 energization confirmation, first Microsoft revenue recognition, Prince George ARR announcement, AI cloud as % of revenue crossing 25%.

Scuttlebutt Findings

• Analyst consensus (constructive): 10 Buy / 2 Hold / 2 Sell; median PT 80— 12036 current price. Post-miss PT cuts were modest; bull thesis broadly intact. (Scuttlebutt stage, 2026-02-23)
• NVIDIA Preferred Partner: Direct relationship signals priority GPU allocation during supply scarcity. Meaningful competitive moat vs non-preferred operators. (Scuttlebutt stage)
• Microsoft $9.7B TCV: Strongest available customer validation. Hyperscaler procurement requires meeting stringent reliability, security, compliance, and uptime thresholds — IREN has cleared them. Hard to replicate without a signed contract as proof point. (Scuttlebutt stage)
• Dell case study: IREN cited as AI infrastructure reference. Multi-vendor recognition beyond NVIDIA and Microsoft. (Scuttlebutt stage)
• CFO "Verizon data centers" slip: On Q2 FY26 call, CFO inadvertently referenced Verizon in context suggesting an active or concluded data center services engagement. If confirmed, Verizon as second enterprise anchor would materially de-risk customer concentration. (Transcript digest)
• Unnamed multibillion-dollar software contract: Prepared remarks referenced an unnamed deal at enterprise scale. No timeline given. Likely a hyperscaler or Tier-1 enterprise. (Transcript digest)
• Culper short attack (Jul 2024) — largely disproved: Culper accused IREN of inflated hash rates and capacity claims. Microsoft $9.7B TCV signed 13 months later is the most authoritative possible response. Girard Sharp securities fraud investigation running — headline risk but not a fundamental concern given Microsoft counterparty. (Scuttlebutt stage)
• Insider selling ($66M at peak): Roberts brothers sold significant stakes near highs. Not automatically bearish (founder diversification), but combined with Culper, sustains credibility discount. Maintain amber flag. (Scuttlebutt stage)
• Employee sentiment: Glassdoor sample is small. Travel demands cited (dispersed global sites). No pattern of management failure or cultural dysfunction. Expansion mode, no layoffs. (Scuttlebutt stage)
• CoreWeave comp: Primary private-market comparable at est. ~35BEVforsimilarAIGPUcloudinfrastructure.IRENat 13B EV with $9.7B contracted revenue appears cheap on this comp. CoreWeave IPO expected CY26 — will be a valuation anchor event for the sector. (Scuttlebutt stage)

Valuation Context

The right frame: IREN must be valued on forward ARR, not TTM revenue. The business being reported today is a declining Bitcoin mining operation funding the build-out of a contracted AI cloud business. TTM multiples are noise.

CoreWeave valuation anchor: CoreWeave at est. 35BEVfor 2B+ annual revenue implies ~17x forward EV/Revenue. IREN at ~$13B EV with 2.3BcontractedARRimplies 5.6x.IfIRENearns10xoncontractedARR, fairvalueis 23B. At 15x, ~$34.5B. Current $13B prices in significant execution risk — rightly so, given Sweetwater dependency.

Sensitivity to timing: Every month Sweetwater energization slips from April 2026 costs ~162MinMicrosoftrevenuerecognition.A3 − monthslip= 486M lost in reported revenue. The stock reacts asymmetrically — downside on slip larger than upside on on-time delivery.

Platform & Secular Position

Secular: IREN operates at the intersection of power scarcity and AI infrastructure demand — the most capacity-constrained layer in the AI stack. CY25 hyperscaler capex: Microsoft $80B, Meta $60-65B, Google $75B. All three guided higher or maintained for CY26. Power availability is the binding constraint on GPU deployment globally. IREN owns 4.5 GW of secured power — a scarce commodity taking years and hundreds of millions to replicate.

Platform: Owner-operated infrastructure (not a reseller) with vertically integrated power → infrastructure build → GPU procurement → operations. Full margin capture at the GPU layer vs cloud resellers who pay pass-through markups. The "mullet" era (BTC cash flow funding AI build) is ending — the business is pivoting fully to AI cloud. Power portfolio extensibility: 4.5 GW secured enables scaling well beyond $3.4B ARR without additional land/power acquisition.

TAM penetration: 810 MW operating / 4.5 GW secured = 18% of secured power capacity deployed. At the $3.4B ARR target (end CY26), still consuming only ~10% of secured power. Power moat supports 10x+ long-run scaling from current operations with no incremental land acquisition required.

Key Risks

1. Sweetwater energization timing. Microsoft revenue commencement April 2026 is entirely dependent on Sweetwater 1 energizing on schedule. Any interconnection delay, equipment failure, or permitting issue pushes $1.94B/yr ARR to the right. Single highest-probability near-term risk.

2. Bitcoin mining cliff. BTC mining fell 28% QoQ in Q2 FY26 and is declining structurally. Another -30% in Q3 FY26 before Microsoft revenues flow could push total revenue below $140M. Psychologically damaging even if the underlying thesis is intact.

3. Dilution trajectory. 75%+ share count growth in 15 months. GPU financing covers most GPU capex, but scaling to $9.2B+ of infrastructure and Oklahoma site buildout (CY28+) may require additional equity. Every capital raise reprices the thesis.

4. Insider selling + litigation. $66M in founder sales near peak plus ongoing Girard Sharp investigation creates persistent credibility drag. Not a fundamental concern given Microsoft counterparty, but suppresses multiple in an already-skeptical market.

5. Hyperscaler capex sentiment shift. AI infrastructure valuations are entirely dependent on continued hyperscaler build-out. If Microsoft or another anchor signals slowdown (macro recession, AI ROI concerns, antitrust action), the sector reprices immediately. Current signals are positive; the risk is binary and exogenous.

Key Catalysts

1. Q3 FY26 earnings (~May 2026): First Microsoft revenues. The single most important near-term catalyst. Even partial-quarter contribution at $1.94B/yr would transform reported revenue and validate the thesis. If Sweetwater energizes April 2026, Q3 FY26 revenue could jump to $350-400M+ from $184.7M.

2. Prince George ARR >$0.5B announcement (Mar-Apr 2026). Management guided contracted ARR to cross $0.5B in "coming weeks" from the Feb 5 call. Formal announcement adds incremental confidence ahead of Q3 FY26.

3. Additional customer announcement. "Verizon data centers" CFO slip + unnamed "multibillion-dollar software contract" suggests a second anchor customer in the pipeline. Announcement would materially de-risk customer concentration.

4. CoreWeave IPO (CY26). CoreWeave going public at a high comparable multiple provides a valuation anchor for IREN. $35B+ for CoreWeave makes $13B EV for IREN look cheap on a comparable basis.

5. 140k GPU deployment and $3.4B ARR achievement (end CY26). Validates end-to-end execution capability at scale. Step-change re-rating catalyst if achieved on schedule.

Position: Atlas holds no position. wsm007 holds IREN at 5.7% of portfolio.