IREN Limited
IREN is building a real AI infrastructure business, but at $41.23 the stock already prices in a near-flawless ramp while earnings, cash flow, and delivery execution remain far behind the valuation.
IREN is no longer just a bitcoin miner. It is trying to become an AI infrastructure platform, and the raw ingredients are real: $3.1B contracted ARR, a $3.4B / 5-year NVIDIA contract, Microsoft deployment progress, 5 GW secured power, and a statement from management that all operating GPU capacity is contracted. That is enough to justify investor interest. It is not enough to justify paying almost any price.
At $41.23, investors are underwriting a very specific outcome. They are assuming the company reaches 480 MW of AI cloud capacity, 150,000 GPUs, and $3.7B ARR by the end of calendar 2026, with limited slippage and attractive margins. Yet the current income statement still does not support that confidence. FY2025 revenue was only $0.5B, operating margin was -13.7%, net margin was -18.1%, and free cash flow was $-0.9B. The latest quarter showed why the market is torn: AI cloud revenue nearly doubled to $33.6M, but total revenue still fell to $144.8M from $184.7M q/q, mining revenue dropped by more than $56M q/q, and net loss reached $247.8M.
The core mispricing is not that the market ignores AI. It is the opposite. The market is capitalizing the future AI story while underweighting the path dependence required to get there. Management itself said the ARR ramp is back-end weighted, with Microsoft revenue and the additional 50,000 GPUs expected to begin ramping in Q3 2026. That means every construction milestone, commissioning handoff, and customer acceptance event matters. In a stock with 30.34% short interest, that can create violent upside squeezes, but it also means a single delayed quarter can force another sharp reset.
What the case depends on is straightforward. First, the Microsoft Horizons rollout must land on schedule. Second, the company must convert contracted capacity into recognized revenue and cash, not just ARR headlines. Third, leverage and capex must stay contained enough that the AI business scales before the balance sheet gets stretched further. Those outcomes are possible, which is why conviction is low rather than high. But with 1030.75x P/E and 11.8x sales, the stock still leaves too little room for error.
The 150,000 GPU and $3.7B ARR targets are hit on time, revenue inflects sharply in late 2026, and heavy short interest fuels an overshoot.
AI ARR grows, but revenue recognition and margins lag the narrative, leaving valuation to compress toward the engine target despite business progress.
Microsoft and NVIDIA deployments slip, impairments continue, and the market rerates the stock closer to infrastructure peers as negative FCF persists.
Quant call (SELL) diverges from analyst consensus (Strong Buy: 7 buy / 0 hold / 1 sell) — resolve explicitly.
The four most dangerous words in investing are: 'this time it's different.'
Forward earnings power is the whole debate here, because trailing numbers are still distorted by the mining-to-AI transition, but even on a forward basis the stock is being valued as if the CY2026 buildout converts into durable high-margin revenue with minimal delay. The moat is better than it was two years ago, built around 5 GW secured power, customer prepayments, and large-scale NVIDIA and Microsoft relationships, but 3 to 5 years out this is still an infrastructure business where returns can be competed down if capital keeps flooding into AI data centers. AI is clearly a tailwind for demand, and IREN is one of the few small-cap names with real contracted capacity, but the secular tailwind does not erase execution risk, financing needs, or the chance that hyperscalers and larger private operators capture most of the economics. At $41.23, the market is paying for the secular story before the earnings stream is proven.
- Valuation is extreme, at 1030.75x P/E versus 18.1x peer average and 11.8x P/S versus 3.0x peers.
- Execution has been weak, with 7 misses in the last 8 quarters, including -238.8% EPS miss on 2026-05-07 and -326.8% on 2025-11-06.
- Capital intensity remains high, with $4.0B debt, 13.88x debt/EBITDA, and $-0.9B FCF as the AI build ramps.
- AI cloud revenue reached $33.6M, nearly doubling from $17.3M q/q, showing the new business is scaling.
- Contracted ARR hit $3.1B and management reiterated $3.7B ARR by CY2026 year-end, with all operational GPU capacity contracted.
- $2.6B cash plus GPU financing and prepayments, including approximately 95% Microsoft-related GPU CapEx funding, reduces near-term funding risk.
- Adjusted EBITDA was $59.5M with a 41.3% margin, a sharp improvement from -123.5% in the prior quarter.
- Quarterly revenue fell to $144.8M from $184.7M q/q during the mining-to-AI transition, showing the current business mix is still in disruption.
- Net loss was $247.8M in the latest quarter, including $140.4M of noncash impairments and $23.7M of unrealized cap-call losses.
- Bitcoin mining revenue fell to $111.2M from $167.4M q/q, and management said additional noncash impairments are still expected as hardware is decommissioned.
- Short interest is 30.34%, with 93.7M shares short, which reflects unusually high skepticism around the ramp and can amplify downside if another delivery slip hits.
| Metric | Value | Context |
|---|---|---|
| Current price | $41.23 | Reference price for valuation and scenario framing |
| Quant price target | $15.44 | -62.5% vs current, Low confidence, basis: earnings_growth_blend |
| Revenue | $0.5B | FY2025, 67.6% YoY growth |
| EPS | $0.41 | FY2025, 55.6% YoY growth, but valuation remains extreme |
| Operating margin | -13.7% | FY2025, improved from -14.6% prior year |
| Net margin | -18.1% | FY2025 profitability still negative |
| P/E | 1030.75x | Vs 18.1x peer average across 7 comparables |
| P/S | 11.8x | Vs 3.0x peer average |
| Leverage | $4.0B debt | 1.49 debt/equity | 13.88x debt/EBITDA | Balance sheet materially more levered than the 2024 Q4 zero-debt period |
| Free cash flow | $-0.9B | Heavy investment phase, with $1.0B capex and no buybacks or dividend |
| Contracted ARR / CY2026 target | $3.1B / $3.7B | Current contracted ARR and reiterated year-end target from 2026 Q3 call |
| 52-week position | $17.22-$76.87 | Current price is -46.4% vs high and 13.2% below 200-day average of $47.49 |
The biggest risk to the SELL call is that the back-end weighted AI ramp lands on time and converts $3.1B contracted ARR into visible revenue faster than the market expects, forcing a squeeze in a stock with 30.34% short interest.
The next catalyst is earnings on August 27, 2026, when investors need proof that Microsoft handoff, the additional 50,000 GPUs, and the CY2026 480 MW / 150,000 GPU / $3.7B ARR targets remain on schedule.
Positioning is volatile and two-sided. Institutional ownership is not especially concentrated, with the top 3 at 24.6%, insider ownership is 4.2%, and short interest is exceptionally high at 30.34% or 93.7M shares, up 23.3%, which signals deep skepticism but also real squeeze risk. Options positioning is not supportive, with a negative gamma regime and a zero-gamma level at $43.93, above spot, which can worsen downside pressure if the stock stays below that level.
Revenue fell from $167.4M q/q as lower BTC prices and hardware decommissioning weighed during the pivot to AI.
Nearly doubled from $17.3M q/q as Microsoft-related deployments and GPU capacity ramped.
5 GW secured power spans North America, Europe, and APAC pipeline, but revenue by geography was not provided.
IREN reiterated aggressive 2026 targets as contracted ARR hit $3.1B and major NVIDIA/Microsoft deployments advance.
The execution record is mixed at best. On the positive side, management did pivot from a pure mining story to an AI infrastructure story with real contracts, moving from $3.1M FY24 AI cloud revenue to $33.6M in the latest quarter and from FY24 adjusted EBITDA of $54.7M to $59.5M in the latest quarter alone, while broadening the platform with Mirantis and major NVIDIA and Microsoft programs. But the scorecard against forecasts and expectations is poor: management said in 2024 Q4, "We reiterate our original unchanged guidance that we will hit 20 exahash by the end of next month and 30 exahash by the end of this year," yet the business has since shifted away from mining amid decommissioning and expected impairments. Street-facing predictability has been weak, with 7 misses in the last 8 quarters and four especially severe misses of -1105.4%, -326.8%, -238.8%, and -214.3%. Credibility on strategic ambition is decent, credibility on timing and clean delivery is still unproven.
In 2026, we are targeting 480 megawatts of AI cloud capacity, 150,000 GPUs and $3.7 billion of ARR by year-end.
We continue to target $3.7 billion in ARR by the end of calendar 2026. We expect that ramp to be back-end weighted with Microsoft revenue, and revenue from the additional 50,000 GPUs procured during the quarter expected to begin ramping in Q3 2026.
Adjusted EBITDA was $59.5 million, compared to $75.3 million in the prior quarter, primarily on account of the revenue and cost of revenue items noted above.
All of our operational capacity is fully contracted. We are not chasing demand. We are racing to build supply fast enough to meet it.
There are no idle GPUs.
The real moat is access to power, speed of energization, and signed demand. IREN now has 5 GW secured power, a $3.4B / 5-year NVIDIA contract, Microsoft-linked deployments, and management says "All of our operational capacity is fully contracted" and "There are no idle GPUs." That is meaningful today, but durability is only moderate because AI infrastructure is capital-intensive and larger hyperscalers, private equity-backed data center operators, and incumbents can still replicate capacity if returns stay outsized.
Capital allocation is aggressive and growth-first. The company is spending heavily into AI capacity, with $1.0B capex, $-0.9B FCF, no buybacks, and no dividend, which is sensible only if the contracted AI ramp converts into cash generation on schedule. The discipline point is mixed: near-term funding looks better than it did historically because management cited $2.6B cash and said approximately 95% of Microsoft-related GPU CapEx is funded, but leverage has risen sharply to $4.0B debt from zero debt around the 2024 Q4 period.
IREN rates a SELL, Low conviction. The company has assembled enough contracts, power, and counterparties to matter in AI infrastructure, but the stock still prices that future too aggressively relative to current earnings, cash flow, and execution evidence. The right question is not whether the AI story is real, it is whether that story is already more than fully reflected at $41.23. A view change would require proof that the back-half 2026 ramp is converting into recognized revenue, sustained positive cash generation, and a valuation that no longer assumes near-perfect delivery.
≣ Full research note tap to expand
Executive Summary
SELL, Low conviction. IREN rates a SELL, Low conviction. The company has secured meaningful AI demand, with $3.1B contracted ARR and a $3.7B ARR year-end target, but the stock still trades at 1030.75x P/E and 11.8x sales despite repeated EPS misses, $-0.9B FCF, and a back-end weighted deployment schedule. The setup is not a short on business quality alone, it is a short on the gap between a promising forward narrative and a valuation that assumes the ramp lands cleanly.
Investment Thesis
IREN is no longer just a bitcoin miner. It is trying to become an AI infrastructure platform, and the raw ingredients are real: $3.1B contracted ARR, a $3.4B / 5-year NVIDIA contract, Microsoft deployment progress, 5 GW secured power, and a statement from management that all operating GPU capacity is contracted. That is enough to justify investor interest. It is not enough to justify paying almost any price.
At $41.23, investors are underwriting a very specific outcome. They are assuming the company reaches 480 MW of AI cloud capacity, 150,000 GPUs, and $3.7B ARR by the end of calendar 2026, with limited slippage and attractive margins. Yet the current income statement still does not support that confidence. FY2025 revenue was only $0.5B, operating margin was -13.7%, net margin was -18.1%, and free cash flow was $-0.9B. The latest quarter showed why the market is torn: AI cloud revenue nearly doubled to $33.6M, but total revenue still fell to $144.8M from $184.7M q/q, mining revenue dropped by more than $56M q/q, and net loss reached $247.8M.
The core mispricing is not that the market ignores AI. It is the opposite. The market is capitalizing the future AI story while underweighting the path dependence required to get there. Management itself said the ARR ramp is back-end weighted, with Microsoft revenue and the additional 50,000 GPUs expected to begin ramping in Q3 2026. That means every construction milestone, commissioning handoff, and customer acceptance event matters. In a stock with 30.34% short interest, that can create violent upside squeezes, but it also means a single delayed quarter can force another sharp reset.
What the case depends on is straightforward. First, the Microsoft Horizons rollout must land on schedule. Second, the company must convert contracted capacity into recognized revenue and cash, not just ARR headlines. Third, leverage and capex must stay contained enough that the AI business scales before the balance sheet gets stretched further. Those outcomes are possible, which is why conviction is low rather than high. But with 1030.75x P/E and 11.8x sales, the stock still leaves too little room for error.
Key Metrics
| Metric | Value | Context | ||
|---|---|---|---|---|
| Current price | $41.23 | Reference price for valuation and scenario framing | ||
| Quant price target | $15.44 | -62.5% vs current, Low confidence, basis: earnings_growth_blend | ||
| Revenue | $0.5B | FY2025, 67.6% YoY growth | ||
| EPS | $0.41 | FY2025, 55.6% YoY growth, but valuation remains extreme | ||
| Operating margin | -13.7% | FY2025, improved from -14.6% prior year | ||
| Net margin | -18.1% | FY2025 profitability still negative | ||
| P/E | 1030.75x | Vs 18.1x peer average across 7 comparables | ||
| P/S | 11.8x | Vs 3.0x peer average | ||
| Leverage | $4.0B debt | 1.49 debt/equity | 13.88x debt/EBITDA | Balance sheet materially more levered than the 2024 Q4 zero-debt period |
| Free cash flow | $-0.9B | Heavy investment phase, with $1.0B capex and no buybacks or dividend | ||
| Contracted ARR / CY2026 target | $3.1B / $3.7B | Current contracted ARR and reiterated year-end target from 2026 Q3 call | ||
| 52-week position | $17.22-$76.87 | Current price is -46.4% vs high and 13.2% below 200-day average of $47.49 |
Financial Strength
The balance sheet is not distressed near term, but it is no longer clean. Liquidity looks adequate, with a 3.72 current ratio and management citing $2.6B cash plus financing and prepayments to support the GPU build, yet this is offset by $4.0B debt, 13.88x debt/EBITDA, and deeply negative free cash flow. Margin direction is improving in the AI business, shown by the latest 41.3% adjusted EBITDA margin, but consolidated economics remain noisy because the company is still shutting down lower-value mining hardware and taking impairments. This is a financeable transition, not a self-funding one.
Competitive Position
The real moat is access to power, speed of energization, and signed demand. IREN now has 5 GW secured power, a $3.4B / 5-year NVIDIA contract, Microsoft-linked deployments, and management says "All of our operational capacity is fully contracted" and "There are no idle GPUs." That is meaningful today, but durability is only moderate because AI infrastructure is capital-intensive and larger hyperscalers, private equity-backed data center operators, and incumbents can still replicate capacity if returns stay outsized.
Management & Guidance
Management is still guiding aggressively. The headline targets are 480 MW AI cloud capacity, 150,000 GPUs, and $3.7B ARR by CY2026 year-end, with 1,210 MW platform capacity by CY2027 and 730 MW under construction. The key caveat is management's own wording that the ramp is back-end weighted, with Microsoft revenue and the additional 50,000 GPUs only beginning to ramp in Q3 2026. That leaves little room for delivery slippage, and the guidance should be discounted because the earnings track record is weak, with 7 misses in the last 8 quarters despite improving strategic positioning.
Segment Analysis
- Bitcoin mining ($111.2M of latest-quarter revenue): Revenue fell from $167.4M q/q as lower BTC prices and hardware decommissioning weighed during the pivot to AI.
- AI cloud ($33.6M of latest-quarter revenue): Nearly doubled from $17.3M q/q as Microsoft-related deployments and GPU capacity ramped.
- Geography (Not disclosed): 5 GW secured power spans North America, Europe, and APAC pipeline, but revenue by geography was not provided.
Capital Allocation
Capital allocation is aggressive and growth-first. The company is spending heavily into AI capacity, with $1.0B capex, $-0.9B FCF, no buybacks, and no dividend, which is sensible only if the contracted AI ramp converts into cash generation on schedule. The discipline point is mixed: near-term funding looks better than it did historically because management cited $2.6B cash and said approximately 95% of Microsoft-related GPU CapEx is funded, but leverage has risen sharply to $4.0B debt from zero debt around the 2024 Q4 period.
Management Execution & Track Record
The execution record is mixed at best. On the positive side, management did pivot from a pure mining story to an AI infrastructure story with real contracts, moving from $3.1M FY24 AI cloud revenue to $33.6M in the latest quarter and from FY24 adjusted EBITDA of $54.7M to $59.5M in the latest quarter alone, while broadening the platform with Mirantis and major NVIDIA and Microsoft programs. But the scorecard against forecasts and expectations is poor: management said in 2024 Q4, "We reiterate our original unchanged guidance that we will hit 20 exahash by the end of next month and 30 exahash by the end of this year," yet the business has since shifted away from mining amid decommissioning and expected impairments. Street-facing predictability has been weak, with 7 misses in the last 8 quarters and four especially severe misses of -1105.4%, -326.8%, -238.8%, and -214.3%. Credibility on strategic ambition is decent, credibility on timing and clean delivery is still unproven.
Positioning & Flows
Positioning is volatile and two-sided. Institutional ownership is not especially concentrated, with the top 3 at 24.6%, insider ownership is 4.2%, and short interest is exceptionally high at 30.34% or 93.7M shares, up 23.3%, which signals deep skepticism but also real squeeze risk. Options positioning is not supportive, with a negative gamma regime and a zero-gamma level at $43.93, above spot, which can worsen downside pressure if the stock stays below that level.
Bull Case
- AI cloud revenue reached $33.6M, nearly doubling from $17.3M q/q, showing the new business is scaling.
- Contracted ARR hit $3.1B and management reiterated $3.7B ARR by CY2026 year-end, with all operational GPU capacity contracted.
- $2.6B cash plus GPU financing and prepayments, including approximately 95% Microsoft-related GPU CapEx funding, reduces near-term funding risk.
- Adjusted EBITDA was $59.5M with a 41.3% margin, a sharp improvement from -123.5% in the prior quarter.
Bear Case
- Quarterly revenue fell to $144.8M from $184.7M q/q during the mining-to-AI transition, showing the current business mix is still in disruption.
- Net loss was $247.8M in the latest quarter, including $140.4M of noncash impairments and $23.7M of unrealized cap-call losses.
- Bitcoin mining revenue fell to $111.2M from $167.4M q/q, and management said additional noncash impairments are still expected as hardware is decommissioned.
- Short interest is 30.34%, with 93.7M shares short, which reflects unusually high skepticism around the ramp and can amplify downside if another delivery slip hits.
Valuation & Price Target
Engine price target $15.44 (-62.5% vs current), Low confidence.
- P/E Multiple: $1.45 (1% weight)
- PEG (growth-adjusted): $2.17 (20% weight)
- P/S Multiple: $3.68 (11% weight)
- Quality-adjusted: $35.62 (20% weight)
The rating aligns with the quant baseline SELL (Low conviction). The engine's $15.44 target is directionally consistent with the core issue, valuation still runs far ahead of delivered economics, even after the stock fell 46.4% from its $76.87 high. There is a legitimate argument against the quant view, because the quality-adjusted method implies $35.62, the company has $3.1B contracted ARR, $2.6B cash, and a much stronger forward AI profile than trailing earnings show. But that upside case does not outweigh the current facts: 1030.75x P/E, 11.8x P/S, 13.88x debt/EBITDA, $-0.9B FCF, a back-end weighted ramp, and a brutal earnings miss record. Sell-side consensus sits at Strong Buy, 7 buy / 0 hold / 1 sell, but that is a low-signal context point, not a reason to ignore the mismatch between fundamentals delivered and valuation paid.
Risk Assessment
Risk score 62/100 (Moderate). The biggest risk to the SELL call is that the back-end weighted AI ramp lands on time and converts $3.1B contracted ARR into visible revenue faster than the market expects, forcing a squeeze in a stock with 30.34% short interest.
The Bottom Line
IREN rates a SELL, Low conviction. The company has assembled enough contracts, power, and counterparties to matter in AI infrastructure, but the stock still prices that future too aggressively relative to current earnings, cash flow, and execution evidence. The right question is not whether the AI story is real, it is whether that story is already more than fully reflected at $41.23. A view change would require proof that the back-half 2026 ramp is converting into recognized revenue, sustained positive cash generation, and a valuation that no longer assumes near-perfect delivery.
Go deeper in the app
Live charts, options analytics, the full earnings-call breakdown, price alerts, and every other stock — updated in real time.
Get Tickr