Back to Portfolio
IREN

IREN Limited

$48.12 16B market cap April 15, 2026
IREN Limited (formerly Iris Energy) IREN BUFFETT / MUNGER / KLARMAN SUMMARY
1 SNAPSHOT
Price$48.12
Market Cap16B
2 BUSINESS

IREN is the most credible Bitcoin-miner-to-AI-cloud pivot story in public markets, differentiated by a $9.7B Microsoft contract that no competitor has matched at this scale. The secured 4.5 GW of renewable power, NVIDIA Preferred Partner status, and aggressive capital deployment ($9.2B funding secured) demonstrate genuine execution capability. However, at $48/share (33.6x EV/Revenue, 131.8x EV/EBITDA), the stock prices in perfect execution of a business transformation that has barely begun -- AI Cloud is still only 9% of revenue. The business has never generated positive free cash flow, shares have diluted 50% in 6 months, $3.3B in convertible notes create further dilution overhang, and the entire AI thesis rests on a single customer (Microsoft). For value investors, this is a classic "wonderful story, terrible price" situation. The company may indeed become a major AI infrastructure player, but the margin of safety is negative at current levels. Wait for $18-22 or proof of sustainable AI profitability before considering a position.

3 MOAT None-to-Narrow

Power cost arbitrage (<$0.033/kWh renewable), 4.5 GW secured grid capacity, $9.7B Microsoft contract. All replicable by well-capitalized competitors.

4 MANAGEMENT
CEO: Daniel Roberts (Co-Founder)

Aggressive/High-Risk - Secured $9.2B funding (converts, GPU financing, prepayments) to fund massive buildout. Betting the company on AI pivot. No dividends, no buybacks.

5 ECONOMICS
-13.9% Op Margin
6.6% ROIC
22.5% ROE
55.1x P/E
-1.13B FCF
17.1% Debt/EBITDA
6 VALUATION
FCF Yield-7.1%
DCF Range15 - 25

Overvalued by 92-220%. Current price ($48) prices in full bull case. Base case SoP = $22.75/share.

7 MUNGER INVERSION
Kill Event Severity P() E[Loss]
Execution risk on $9.7B Microsoft contract -- requires deploying 140K GPUs, building liquid-cooled DCs, energizing 4.5GW of power on schedule HIGH - -
Bitcoin price volatility -- mining still 91% of revenue; BTC below $40K makes mining unprofitable at IREN's cost structure MED - -
8 KLARMAN LENS
Downside Case

Execution risk on $9.7B Microsoft contract -- requires deploying 140K GPUs, building liquid-cooled DCs, energizing 4.5GW of power on schedule

Why Market Right

GPU delivery delays from Dell/NVIDIA supply chain; Microsoft contract renegotiation or volume reduction; Bitcoin price decline below $50K making mining unprofitable; Further dilutive capital raises (convertible notes conversion, ATM facility); BC Hydro rate increases or regulatory changes

Catalysts

Sweetwater 1 (1,400 MW) energization targeted April 2026 -- imminent capacity catalyst; Microsoft GPU deployment milestones -- each tranche unlocks revenue recognition toward $1.94B annual target; AI Cloud ARR ramp toward $3.4B target by CY26 end; Bitcoin price strength above $75K sustains mining profitability; Second major AI customer announcement would derisk concentration

9 VERDICT WAIT
C+ Quality Moderate-Weak - $3.26B cash war chest offset by $3.69B convertible debt, negative FCF, and $5.8B committed GPU spend. Capital markets dependent.
Strong Buy$12
Buy$22
Fair Value$25

REJECT at current prices. Monitor for entry at $18-22 (>50% decline) or proof of sustained AI Cloud profitability (2 quarters of >$200M AI revenue with >50% gross margins).

🧠 ULTRATHINK Deep Philosophical Analysis

IREN Limited -- Deep Philosophical Analysis

Buffett/Munger/Klarman Style Thinking | April 15, 2026

The Core Question: Are You Buying Infrastructure or Buying a Story?

There is a useful mental exercise for evaluating IREN: imagine you are not buying a stock but instead being offered partnership in a business. Someone comes to you and says, "We run data centers in Texas and British Columbia. We mine Bitcoin today, but we're converting to AI cloud services. We just signed a $9.7 billion contract with Microsoft to host their GPUs. We need about $6 billion more in capital to fulfill it. Want in?"

Your first question should be: "What is my rate of return on that $6 billion?"

And here is where the IREN story gets complicated. The company has secured $9.2 billion in funding -- customer prepayments, convertible notes, GPU leasing, GPU financing -- to deploy 140,000 GPUs generating a targeted $3.4 billion in annual recurring revenue. If the margins are 30-35%, that is roughly $1 billion to $1.2 billion in annual gross profit on $9.2 billion of capital deployed. That is a return on invested capital of 11-13%.

For infrastructure with significant technology obsolescence risk (GPU generations turn over every 18-24 months), execution risk (building liquid-cooled data centers at unprecedented scale), and customer concentration risk (essentially a single counterparty), 11-13% is not compelling. A toll bridge earns 11-13% returns with near-zero risk. IREN is earning toll bridge returns while carrying venture capital risk.

Munger would say: "The trouble with this business is that you are a capital intermediary. You borrow at 6%, buy GPUs, put them in buildings you build, and rent them to Microsoft. You are essentially a bank for compute infrastructure. Banks earn 1% on assets. What makes you think you can do better permanently?"

Moat Meditation: The Illusion of Scarcity

The IREN bull case rests on a single premise: AI compute demand will outstrip supply for the foreseeable future, making anyone with secured power and GPU access a bottleneck owner. This is the Aschenbrenner thesis in a nutshell -- that the "overhang" of AI demand is so large that even mediocre infrastructure providers will earn excess returns simply by existing in the right place at the right time.

But scarcity is temporal, not structural. In 2021, semiconductor fabs were "scarce" and TSMC was building as fast as possible. By 2023, there were gluts in certain nodes. In 2020, shipping containers were "scarce" and freight rates went to $20,000/container. By 2023, rates were back to $1,500. The lesson of every infrastructure boom is that capital flows to scarcity until scarcity is destroyed.

IREN has 4.5 GW of secured power. That sounds enormous until you realize Microsoft alone plans to spend $80 billion on data center capital expenditure in fiscal year 2025. Amazon, Google, Meta, and Oracle are spending comparable amounts. The total addressable power demand for AI data centers is projected at 50-100 GW by 2030. IREN's 4.5 GW is 5-9% of that demand -- meaningful but not monopolistic.

The true moat question is: in five years, when hundreds of gigawatts of data center capacity have been built worldwide, what will IREN have that competitors cannot replicate? The honest answer is: not much. Power contracts can be secured by anyone with capital. Land can be purchased. Data centers can be built. The GPUs belong to NVIDIA and the customer relationship belongs to Microsoft. IREN is a middleman, and middlemen in commodity infrastructure earn commodity returns.

The Owner's Mindset: Would Buffett Own This for 20 Years?

This is the question that ends most speculative infrastructure analyses. Buffett's entire framework is built on identifying businesses whose competitive advantages compound over time -- businesses that are worth more in 20 years not because the stock went up but because the underlying economics improved.

What does IREN look like in 20 years? The honest answer is: we have no idea. The company might be a $50 billion AI infrastructure giant. It might be bankrupt. It might have been acquired by Microsoft or Dell. The range of outcomes is extraordinarily wide, and that width of outcomes is itself a disqualifying factor for the Buffett framework.

The businesses Buffett buys -- Coca-Cola, See's Candies, Apple -- have narrow ranges of outcomes. They will be around in 20 years doing roughly what they do today, a little bigger, a little more profitable. The value compounds slowly, reliably, predictably.

IREN is the opposite. Its outcome depends on: (1) whether AI compute demand grows as projected, (2) whether NVIDIA maintains GPU dominance or new architectures emerge, (3) whether Microsoft maintains its cloud leadership, (4) whether IREN can execute a multi-billion-dollar buildout without cost overruns, (5) whether Bitcoin remains viable, (6) whether Texas and BC power markets remain favorable, and (7) whether capital markets remain willing to fund the operation. That is seven independent variables, each of which must break favorably.

Buffett would say: "I don't invest in things where I need seven things to go right. I invest in things where only one thing needs to go right: the business needs to keep doing what it's already doing."

Risk Inversion: What Could Destroy This Business?

Inverting is Munger's favorite tool. Instead of asking "what could go right?", ask "what would make me lose all my money?"

  1. Microsoft walks away or renegotiates. The $9.7B contract has ramp conditions, performance milestones, and presumably termination clauses. If IREN fails to deliver GPUs on time, Microsoft has contractual remedies. If AI demand cools, Microsoft has no obligation to maintain volume above minimums. This single point of failure could vaporize 90% of the stock's value.

  2. GPU technology shifts. NVIDIA GB300 Blackwell is today's architecture. In 18-24 months, the next generation arrives. IREN's $5.8B investment in current-generation GPUs begins depreciating from day one. If a competitor deploys next-generation hardware at lower cost, IREN's contracted rates become above-market and the Microsoft relationship erodes.

  3. The convertible notes convert at depressed prices. $3.3 billion in convertible notes at potentially low conversion prices could double the share count. In a downturn, this creates a vicious cycle: stock falls, conversion price falls, dilution increases, stock falls further.

  4. Bitcoin bear market combined with AI delays. If BTC falls below $40K (mining turns unprofitable) at the same time GPU deployments are delayed (no AI revenue), IREN has $3.69B in debt, negative cash flow, and no profitable business line. This scenario is not farfetched -- it describes roughly what happened to crypto miners in 2022.

  5. Regulatory action on power consumption. Texas and BC are currently friendly to data center power demands, but public opinion could shift rapidly if energy prices spike or grid reliability declines.

Valuation Philosophy: Paying for Perfection vs. Paying for Protection

Seth Klarman wrote: "Value investing is at its core the marriage of a contrarian streak and a calculator." The calculator on IREN is unforgiving. At $48 per share, investors are paying $16 billion for a business that generates $250M in annual operating cash flow, has $3.69B in debt, and has never produced a dollar of free cash flow.

The entire valuation rests on forward expectations: $3.4B ARR by end of CY2026, expanding margins, eventual FCF generation. These are projections, not facts. And the market is pricing them as facts.

Klarman's margin of safety principle demands that you pay substantially less than your conservative estimate of intrinsic value. Our base case intrinsic value is $22.75 per share. At $48, investors have negative margin of safety -- they are paying 111% above fair value. This is the opposite of value investing.

The only scenario where $48 makes sense is the bull case ($49.50/share), which requires full, on-time GPU deployment, $3.4B ARR achievement, 35% margins, and no further dilution. The probability-weighted expected value, using 20% bull / 50% base / 30% bear, is approximately $21/share. The stock is priced at 2.3x its expected value.

The Patient Investor's Path

For the disciplined value investor, IREN presents a clear framework:

Today ($48): Do nothing. The risk-reward is unfavorable. You are paying bull-case prices for a business with bear-case risks. Let others take the execution risk.

At $30-35 (30-40% decline): Watch closely. At this level, the base case begins to look reasonable. Monitor Q3-Q4 FY26 AI revenue ramp. If AI Cloud revenue exceeds $200M/quarter with >40% margins, the story is becoming real.

At $18-22 (55-65% decline): Consider a small position. This is the accumulate zone where the mining business alone provides support ($1.4-2.3B mining value / 400M FD shares = $3.50-5.75/share) and the AI optionality comes nearly free. Size it at 1-2% maximum.

At $10-12 (75%+ decline): Strong buy territory. The mining business covers the price, the AI business is free optionality, and the Microsoft contract provides a floor. This is where Klarman or Buffett would get interested -- buying a real asset (4.5 GW of power infrastructure) for less than replacement cost.

The hardest thing in investing is watching a stock you admire from the wrong side of the price equation. IREN may indeed be a transformative AI infrastructure company. But transformation at any price is not investing -- it is speculation. Wait for the price to reflect the risks, not just the dreams.

"The stock market is a device for transferring money from the impatient to the patient." -- Warren Buffett

In the case of IREN, the patient investor's job is to wait for the price to reflect the business as it is today -- a cash-burning, debt-laden, single-customer infrastructure play -- rather than the business it hopes to become.

Executive Summary

IREN Limited (formerly Iris Energy) is an Australian-headquartered, vertically integrated data center operator that runs two businesses: Bitcoin mining and AI/GPU cloud services. The company is in the middle of a massive pivot from crypto mining to hyperscale AI infrastructure, anchored by a landmark $9.7 billion, five-year contract with Microsoft signed in November 2025. IREN has secured over 4.5 GW of grid-connected power across Texas, Oklahoma, and British Columbia, and targets 140,000 GPU deployments generating $3.4 billion in annualized recurring revenue by end of CY2026.

Leopold Aschenbrenner's Situational Awareness LP holds IREN as its #5 position (7.7% of portfolio), viewing it as a leveraged play on AI infrastructure demand -- the same thesis behind his positions in CORZ (9.8%), CRWV (18.2%), and APLD (6.5%).

Investment Thesis: IREN is a capital-intensive, speculative infrastructure play undergoing a high-stakes transformation from commodity Bitcoin mining to contracted AI cloud services. The $9.7B Microsoft deal is genuinely transformative and differentiates IREN from pure-play miners. However, the stock at $48 prices in near-perfect execution of a multi-year buildout that requires $5.8B in GPU procurement (Dell), $3.6B in financing, and flawless deployment of 140,000 GPUs -- all while burning cash and diluting shareholders. The business has no durable moat in the Buffett/Munger sense: power contracts are replicable, data center colocation is commodity infrastructure, and the Microsoft relationship could be brought in-house or renegotiated.

Recommendation: REJECT for value investing portfolio. WAIT at $18-22 for speculative allocation only. The risk-reward at $48 is unfavorable. Fair value range is $15-25 on base case assumptions.

1. Business Overview

What Does IREN Do?

IREN operates large-scale, renewable-energy-powered data centers across North America serving two revenue streams:

  1. Bitcoin Mining (91% of Q2 FY26 revenue) -- Self-mining Bitcoin using proprietary infrastructure powered by renewable energy. Achieved 50+ EH/s hashrate by mid-2025, representing approximately 5% of global network capacity.

  2. AI Cloud Services (9% of Q2 FY26 revenue, rapidly scaling) -- GPU-as-a-service for hyperscale customers, anchored by the $9.7B Microsoft contract. Deploying NVIDIA GB300 Blackwell GPUs in liquid-cooled data centers.

Data Center Portfolio

Location Capacity Status Primary Use
Childress, Texas 750 MW Operational, expanding Microsoft AI (Horizon 1-4, 200MW IT load)
Sweetwater 1, Texas 1,400 MW Energization April 2026 AI + Mining
Sweetwater 2, Texas 600 MW Energization 2027 AI + Mining
Oklahoma 1,600 MW New campus, 2,000 acres AI infrastructure
Prince George, BC 50+ MW Operational AI Cloud (20,000+ GPUs)
Mackenzie, BC 80 MW Operational Mining + AI
Canal Flats, BC 30 MW Operational Mining
Total Secured >4.5 GW

Revenue Trajectory

Period Total Revenue BTC Mining AI Cloud AI %
FY2023 $75.5M ~$73M ~$2.5M 3%
FY2024 $187.2M $184.1M $3.1M 2%
FY2025 $501.0M $484.6M $16.4M 3%
Q1 FY26 $240.3M $233.0M $7.3M 3%
Q2 FY26 $184.7M $167.4M $17.3M 9%
Target CY26 exit $3.4B ARR

The AI cloud revenue doubled sequentially in Q2 ($17.3M vs $7.3M) even as total revenue fell 23% due to lower Bitcoin mining output. This is the inflection point the bulls are watching.

2. Phase 1: Risk Analysis

2.1 Business Model Risk -- HIGH

Bitcoin Mining Volatility: Q2 FY26 exposed the fundamental problem. Mining revenue fell from $233M to $167M (-28%) in a single quarter. Bitcoin price fluctuations, network difficulty increases (halving cycles), and energy cost variability create unpredictable revenue. At BTC ~$75,000 (current), mining is profitable at IREN's ~$39K/BTC electricity cost, but a 50% BTC price decline to $37,500 would make mining unprofitable.

AI Pivot Execution Risk: The $9.7B Microsoft contract requires IREN to procure $5.8B in NVIDIA GB300 GPUs from Dell, build liquid-cooled data centers, and deploy 140,000 GPUs by end of CY2026. This is a staggering operational challenge for a 457-employee company that was primarily a crypto miner 18 months ago. Any delays in GPU delivery (supply chain), construction, or power energization directly impact revenue recognition.

Capital Market Dependency: IREN cannot self-fund. Q2 FY26 saw $71.6M operating cash flow vs $850.9M capital expenditure. The $3.0B in Q2 financing (convertible notes, prepayments) was necessary to keep the lights on. The business model requires continuous access to capital markets -- a dangerous dependency if credit conditions tighten or sentiment shifts.

2.2 Financial Risk -- HIGH

Risk Factor Current State Severity
Leverage $3.69B convertible notes / $2.51B equity = 1.53x D/E HIGH
Interest Burden $3.6B GPU financing at <6% = ~$216M annual interest HIGH
FCF -$779M Q2 FY26 annualized CRITICAL
Dilution Shares ~332M (up from 221M in 6 months) HIGH
Convertible Notes $3.3B in converts = massive potential dilution HIGH
Derivative Losses -$219M unrealized losses in Q2 MEDIUM

The convertible notes are particularly dangerous. At $3.3B face value, if converted to equity at even $50/share, that adds ~66M shares (20% dilution). If the stock falls, conversion prices adjust downward, creating a dilution death spiral.

2.3 Regulatory & Operational Risk -- MEDIUM

  • BC Hydro dependency: British Columbia operations rely on hydroelectric power contracts. Provincial regulatory changes, drought conditions, or rate increases could impact margins.
  • ERCOT grid risk: Texas operations on ERCOT grid face extreme weather risk (see 2021 Winter Storm Uri). Power price spikes during grid stress events could create massive cost overruns.
  • Cryptocurrency regulation: Potential energy consumption regulations on Bitcoin mining.
  • GPU supply chain: NVIDIA GB300 allocation depends on Dell fulfilling the $5.8B order.

2.4 FX & Governance Risk -- LOW-MEDIUM

  • AUD/USD exposure: Headquartered in Australia, reports in USD. Manageable.
  • Insider ownership: 9.94% -- decent for a growth company. Co-founders Daniel and William Roberts maintain involvement.
  • Short interest: 18.7% of float -- significant bearish sentiment. Short squeeze potential exists but also signals fundamental concerns.

Risk Score: 7.5/10 (High)

3. Phase 2: Financial Analysis

3.1 Profitability Assessment

Gross Margin Volatility is the Red Flag:

Period Gross Margin Comment
FY2022 87.4% Small scale, high BTC prices
FY2023 47.8% BTC bear market
FY2024 53.5% Recovery
FY2025 68.3% BTC bull + scale
Q1 FY26 62.9% Strong
Q2 FY26 64.4% Stable (ex-depreciation)

Adjusted EBITDA tells a different story: Q2 Adjusted EBITDA was $75.3M (41% margin), suggesting the underlying cash-generating ability is better than GAAP numbers show. The gap between GAAP and adjusted is driven by non-cash items: $219M derivative losses, $31.8M mining hardware impairment, $111.8M debt conversion inducement, and $58.2M stock-based comp.

3.2 Cash Flow Analysis

Period Operating CF CapEx FCF Financing
FY2023 $8M -$120M -$112M --
FY2024 $50M -$480M -$430M --
FY2025 $246M -$1,370M -$1,124M --
Q1 FY26 $142M -$281M -$139M $606M
Q2 FY26 $72M -$851M -$779M $3,008M

The math is clear: IREN generates positive operating cash flow ($72-142M/quarter) but spends 4-12x that on capital expenditures. Over the past 18 months, IREN has spent ~$2.5B on capex while generating ~$460M in operating cash flow -- a $2B+ funding gap filled entirely by debt and equity.

3.3 Balance Sheet

Metric FY2024 FY2025 Q2 FY26 Trend
Total Assets $1.15B $2.94B $7.03B Explosive growth
Cash $0.4B $0.6B $3.26B War chest
PP&E ~$0.6B ~$1.8B $3.17B GPU + DC investment
Total Debt $0.06B $1.0B $3.69B Concerning
Equity $1.1B $1.8B $2.51B Growing but diluted
D/E Ratio 0.05x 0.62x 1.53x Deteriorating
Retained Earnings -- -- -$367M Cumulative losses

The $3.26B cash position provides ~4 quarters of runway at current capex rates ($850M/quarter), but the company has committed to $5.8B in GPU purchases from Dell. Without additional financing or Microsoft prepayments ramping, another capital raise is likely by mid-FY2027.

3.4 Valuation

Metric Value Context
Market Cap $16.0B
EV ~$19.4B
P/E (TTM) 55.1x Volatile earnings base
Forward P/E 64.5x Assumes profitability improvement
P/B 9.53x Heavy premium to book
P/S (TTM) 32.8x Extreme premium
EV/EBITDA 131.8x Unsustainable
EV/Revenue 33.6x Priced for hypergrowth
PEG 0.85 Looks cheap IF growth materializes
FCF Yield Negative No free cash flow
Beta 4.32 Extreme volatility

Valuation Conclusion: At 33.6x EV/Revenue with negative free cash flow, IREN is priced for the AI story to fully materialize. Comparable data center REITs (Equinix, Digital Realty) trade at 10-14x EV/Revenue with proven recurring revenue and positive FCF. Even high-growth cloud infrastructure names were priced at 15-20x forward revenue at IPO. IREN's premium is justified ONLY if the $3.4B ARR target is achieved on schedule with margins above 40%.

4. Phase 3: Moat Analysis

4.1 Does IREN Have a Moat?

Moat Source Assessment Durability Score
Power Contracts 4.5 GW secured across TX/OK/BC Medium -- 5-10 years 5/10
Renewable Energy 100% renewable, <$0.033/kWh Low -- table stakes for new DCs 3/10
Microsoft Relationship $9.7B contract, 5-year term Medium -- single customer risk 5/10
Site Control 2,000+ acres Sweetwater, 576 acres Childress Medium -- land is replicable 4/10
Vertical Integration Own substations, cooling, infrastructure Low-Medium -- capex barrier only 4/10
Scale 50+ EH/s mining, 140K GPU target Low -- competitors larger 3/10
NVIDIA Partnership Preferred Partner status Low -- NVIDIA has many partners 3/10

Overall Moat: NARROW-TO-NONE (3/10)

4.2 What IREN Has vs. What a Moat Requires

Power cost arbitrage is real but temporary. IREN's <$0.033/kWh electricity cost (largely hydroelectric in BC, wind/solar in Texas) provides genuine unit economics advantage. Mining one Bitcoin costs ~$39,000 in electricity vs. BTC price of ~$75,000. But this advantage is: (a) dependent on utility contracts that expire, (b) replicable by competitors securing similar contracts, and (c) eroding as renewable energy becomes ubiquitous.

The Microsoft contract is the closest thing to a moat. A $9.7B, 5-year commitment from the world's largest cloud company provides contracted revenue visibility that no pure-play miner can match. The 20% prepayment ($1.9B) plus GPU financing at <6% covers 95% of GPU-related capex. This is meaningful. But Microsoft could renegotiate terms, build its own facilities, or reduce orders -- and IREN's entire AI story collapses without this single customer.

No switching costs, no network effects, no IP. Data center colocation is fundamentally commodity infrastructure. The GPUs are NVIDIA's. The software stack is the customer's. The power comes from the utility. IREN provides the building, cooling, and connectivity -- all replicable.

4.3 Competitive Landscape

Competitor Power Capacity AI Customer Advantage Over IREN
CoreWeave (CRWV) ~1 GW Microsoft, others Pure-play AI cloud, larger GPU fleet
Core Scientific (CORZ) 920 MW CoreWeave More sites, 10 data centers
Marathon Digital (MARA) >1 GW Pivoting to AI Larger mining operation
Riot Platforms (RIOT) 1 GW+ Pivoting to AI Established Texas operations
Applied Digital (APLD) 400+ MW Various Dedicated AI data centers
Hyperscalers (MSFT/AMZN/GOOG) 10+ GW each Internal Could absorb IREN's function

5. Phase 4: Synthesis

5.1 Sum-of-Parts Valuation

Part 1: Bitcoin Mining Business

Assumption Value
Hashrate 50 EH/s
Annual BTC mined (est.) ~6,000-7,000 BTC
BTC price $75,000
Mining revenue ~$475M
Mining gross margin 60%
Mining gross profit ~$285M
Appropriate EV/Gross Profit multiple 5-8x (commodity, cyclical)
Mining business value $1.4B - $2.3B

Part 2: AI Cloud Services Business

Scenario CY26 Exit ARR Margin Earnings Power Multiple Value
Bull (on-time, full ramp) $3.4B 35% $1.19B 15x $17.9B
Base (delays, 70% ramp) $2.0B 30% $600M 12x $7.2B
Bear (major delays, 40% ramp) $1.0B 20% $200M 8x $1.6B

Part 3: Net Debt / Dilution Adjustment

  • Net debt: $3.69B debt - $3.26B cash = $430M
  • Convertible dilution: $3.3B converts could add 50-100M shares
  • Fully diluted shares: ~400M

Sum-of-Parts Fair Value:

Scenario Mining + AI Less Net Debt Per Share (400M FD) vs Current
Bull $20.2B $19.8B $49.50 +3%
Base $9.5B $9.1B $22.75 -53%
Bear $3.9B $3.5B $8.75 -82%

5.2 Entry Price Calculations

Level Price EV/Rev (FY27E) Rationale
Strong Buy $12.00 ~3x Deep value, prices in no AI success
Accumulate $22.00 ~6x Fair value with margin of safety
Fair Value $30.00 ~8x Base case, moderate execution
Current $48.12 ~13x Bull case fully priced
Overvalued $65+ ~17x Requires perfect execution + expansion

5.3 What Would Change Our View

Upgrade triggers (to ACCUMULATE):

  1. Two consecutive quarters of AI Cloud revenue >$200M with gross margins >50%
  2. Demonstrable path to positive free cash flow (even one quarter)
  3. Share count stabilization (no new dilutive issuances for 2+ quarters)
  4. Second major AI customer beyond Microsoft
  5. Stock price falls to $18-22 range

Downgrade triggers (to STRONG SELL):

  1. Microsoft contract renegotiation or delay
  2. GPU delivery delays from Dell/NVIDIA beyond 6 months
  3. BTC price sustained below $50,000
  4. Additional dilutive capital raise above 50M shares
  5. BC Hydro regulatory adverse action

6. Aschenbrenner / SALP Context

Leopold Aschenbrenner's Situational Awareness LP holds IREN at 7.7% of portfolio (~$330M position). This is his 5th largest position behind BE (20.6%), CRWV (18.2%), LITE (11.2%), and CORZ (9.8%).

What Aschenbrenner sees: IREN as a leveraged bet on AI infrastructure scarcity. His essay "Situational Awareness" argues AI compute demand will grow 100-1000x, making power-secured data center operators the critical infrastructure layer. IREN's 4.5 GW of secured power and Microsoft contract fit this thesis perfectly.

What we see differently: This is a venture-style bet, not a value investing thesis. Every other BTC-to-AI pivot play in the SALP portfolio that we have analyzed (CORZ, APLD, CLSK, HUT, BTDR, BITF) has been REJECTED for the same structural reasons: no moat, negative FCF, chronic dilution, commodity infrastructure.

7. Verdict & Recommendation

Rating: REJECT (for value portfolio) / WAIT at $18-22 (for speculative allocation)

Position Sizing: Zero allocation at current prices.

Price Targets:

  • Strong Buy: $12.00
  • Accumulate: $22.00
  • Hold Zone: $22-35
  • Sell Zone: $40+ (current levels)

What IREN Gets Right

  • The Microsoft contract is genuinely transformative. $9.7B over 5 years with 20% prepayment is tier-1 validation.
  • Power infrastructure is real and scarce. 4.5 GW of secured, grid-connected power in renewable-rich locations.
  • Management is executing aggressively. $9.2B in secured funding demonstrates capital markets credibility.

What Keeps Us Out

  • No durable moat. Power contracts expire, data centers are commodity infrastructure, Microsoft has all the leverage.
  • Negative free cash flow. The business has never generated positive FCF and won't for years.
  • Extreme dilution. Shares have increased ~50% in 6 months. $3.3B in converts could add another 50-100M shares.
  • Single customer concentration. The entire AI bull case rests on Microsoft.
  • Valuation assumes perfection. At $48, investors pay for the bull case with zero margin of safety.

Citations

  1. IREN Q2 FY26 Earnings Release (February 5, 2026) -- GlobeNewsWire
  2. Microsoft $9.7B Contract Announcement (November 3, 2025) -- Reuters
  3. AlphaVantage Financial Statements: income-statement.json, balance-sheet.json, cash-flow.json
  4. FinViz Real-Time Quote Data (April 17, 2026)
  5. CoinLaw IREN Statistics 2026
  6. IREN Investor Relations -- iren.com/investors
  7. IREN Data Centers -- iren.com/data-centers

Analysis Date: April 15, 2026 Framework: /Users/fried/Desktop/stockresearch/research/analysis-framework.md