IREN has landed some very large, multi year computing rental contracts with well known technology companies, most notably Microsoft, and that gives it a real and growing base of future revenue.
Building the facilities to deliver on those contracts costs an enormous amount of money, more than a billion dollars a quarter recently, and the company has been raising that money partly by selling a lot of new shares, which dilutes existing shareholders, and one outside estimate suggests it still needs many billions more before the full build out is fully funded.
In early July 2026 the stock also came under pressure after Meta said it plans to rent out its own spare computing power to other companies, a new competitor for exactly the kind of business IREN is chasing, and after IREN's board approved very large stock awards for its two co-CEOs around the same time as a costly new sports sponsorship.
Putting the size of the contracted computing rental business against the size of the ongoing spending, dilution, and new competition it now faces, the stock looks fairly valued right now, neither a clear bargain nor a clear overpay.
A quick look at how a few well known investing styles might weigh this stock.
Investors who follow Cathie Wood's approach would frame IREN as a company converting Bitcoin mining infrastructure into a platform for the artificial intelligence computing buildout, pointing to AI cloud revenue that grew sharply quarter over quarter and more than three billion dollars of future revenue already under contract with Nvidia and Microsoft. A Cathie Wood style investor would weigh the size of that future opportunity more heavily than the company's current profits.
A Chase Coleman style investor would focus on the scale and duration of the Microsoft and Nvidia agreements and how quickly AI cloud revenue is scaling from a small base, treating IREN as a bet on becoming a long term supplier of computing capacity to the largest AI companies. They would watch closely whether the contracted revenue actually converts into growing cash generation as the buildout continues.
Investors who follow Stanley Druckenmiller's approach would see IREN as a way to ride the broader artificial intelligence infrastructure buildout, pointing to the rapid shift in the company's revenue mix away from Bitcoin mining and toward AI cloud contracts. They would size any position based on how the stock and the wider AI infrastructure theme are trading, and would be quick to step back if that momentum turned.
A Bill Ackman style investor would want to pick apart the balance sheet before committing capital, focusing on how the company is funding its data center expansion through convertible debt and share issuance while cash flow from operations still trails the money being spent on new facilities. They would want confidence that the Microsoft and Nvidia contracts turn into durable, growing profit before building a large position.
Investors who follow Warren Buffett's approach would likely stay on the sidelines, noting that IREN has only a short track record as an AI cloud provider and still depends on heavy ongoing capital spending and debt financing to build out its data centers. That combination falls short of the long, predictable run of profit and free cash flow this style of investor typically looks for before buying.
These takes are illustrations of well known investing styles applied to this stock. They are informed guesses, not the actual views, statements, or holdings of the people named.
IREN earns money two ways: operating specialized machines that mine and sell Bitcoin, and leasing graphics chip computing capacity inside its data centers to AI companies under long term cloud contracts, including multi billion dollar agreements with Microsoft and Nvidia. It is steadily shifting power and building space away from Bitcoin mining and toward the AI cloud business as those contracts ramp up, with AI cloud revenue growing far faster than Bitcoin mining revenue in the most recent quarter.
We'd revisit this call if the price falls below $16.00 or rises above $65.00.
These are review triggers for us to take another look, not price targets or a promise the stock will reach them.
Peer benchmark data not available.
** IREN Limited, historically a bitcoin miner, now builds and operates data center capacity it leases to AI computing customers, alongside a shrinking bitcoin mining business. Fiscal year ends June 30. FY2025 (ended June 30, 2025) revenue was $501.0 million, up 168% year over year, with net income of $86.9 million (10-K filed August 28, 2025). Trailing twelve month revenue is approximately $757 million as of the most recent reported quarter.
** Q3 FY26 (ended March 31, 2026) revenue was $144.8 million, down from $184.7 million in Q2, as bitcoin mining revenue fell to $111.2 million (from $167.4 million) on hardware decommissioning, while AI Cloud revenue rose to $33.6 million (from $17.3 million), reaching 23.2% of revenue versus 9.4% the prior quarter. Q3 net loss was $247.8 million (versus a $155.4 million loss in Q2), driven by depreciation and a $140.4 million impairment on retired mining and data center equipment. Nine month FY26 results were flattered by a large non cash unrealized gain on financial instruments (warrant and convertible note fair value marks), masking a much weaker underlying operating result.
** Q3 FY26 operating cash flow was $75.3 million against roughly $1.36 billion of property, equipment, and computer hardware capital expenditure in the same quarter (total investing outflow was $1.48 billion), implying a free cash outflow of about $1.28 billion in that quarter alone.
** As of March 31, 2026: cash and equivalents $2,213.3 million (down from $3,260.6 million at December 31, 2025), convertible notes payable $3,687.8 million, stockholders equity $2,664.5 million. On May 14, 2026 the company closed an additional $3.0 billion of 1.00% convertible senior notes due 2033, adding both cash and leverage. One recent analyst estimate puts IREN's remaining unfunded capital need for its full disclosed build-out plan at roughly $21 billion, against a sector-wide financing need estimated near $50 billion; this is a single third party estimate, not a company disclosure, and should be treated as directional rather than precise.
** Shares outstanding have grown from 223,245,651 per the FY2025 10-K to roughly 357 million as of July 2026, a rise of about 60%, through a December 2025 registered direct offering, an at the market program, and RSU vesting. Contracted AI Cloud annualized revenue is $3.1 billion ($1.9 billion Microsoft, $0.7 billion Nvidia, $0.5 billion a Prince George facility), targeting $3.7 billion by the end of calendar 2026, though the company states the $3.7 billion figure is not fully contracted and actual revenue may differ materially. Short interest is elevated, around 18% of shares outstanding as of early July 2026, contributing to sharp swings (a roughly 41% one month decline followed by a 15% single day squeeze on July 6, 2026).
** TeraWulf has about $12.8 billion in contracted HPC revenue plus a roughly $19 billion, 20 year Anthropic linked lease announced July 6, 2026; Core Scientific has roughly $10 to 12 billion contracted through CoreWeave; Hut 8 signed a $7 billion, 15 year lease; CoreWeave and Nebius separately report order books near $99 billion and $50 billion. IREN's backlog is meaningful but smaller in absolute terms than several of these peers, and one analyst flagged IREN as still catching up on diversifying its customer base beyond Microsoft.
** Bitcoin has fallen from a peak near $126,000 in October 2025 to roughly $63,000 to $64,000 in early July 2026, pressuring the legacy mining segment. Hyperscalers are projected to spend roughly $725 billion on AI infrastructure in 2026, a structural tailwind, but on June 30 to July 1, 2026 Meta disclosed plans (internally called Meta Compute) to resell its own excess AI computing capacity to third parties, which markets read as a new, well capitalized competitor to independent AI cloud providers like IREN; the stock fell roughly 23% over the following five sessions.
** On June 30, 2026 the board approved RSU grants of 9,099,328 shares each (roughly $700 to $788 million combined at recent prices) to co-CEOs William and Daniel Roberts, vesting over four years with an additional two year post-vesting holding period per tranche (through fiscal 2033); the company states neither will receive another equity grant until fiscal 2031. These are compensation grants, not open market purchases, but the size drew investor criticism and coincided with a nine session losing streak alongside a newly announced $50 million a year, multi year Golden State Warriors jersey sponsorship, the largest in North American sports history. No open market insider buying was found.
** Near term: Q4/full year FY26 results expected late August 2026, further capacity and contract milestones. Medium term: full commissioning of the Microsoft (200 megawatt) and Nvidia (60 megawatt) deployments at Childress by 2027. Key risks: Meta's compute resale plans compressing the pricing and growth outlook for independent AI cloud providers; a disclosed multi billion dollar gap between build-out ambitions and secured financing; continued dilution; customer concentration in Microsoft and Nvidia; and further bitcoin price weakness.
** The $140.4 million equipment impairment shows the legacy mining segment eroding faster than a smooth pivot narrative suggests. Recent profitability has been substantially a non cash accounting artifact, not operating cash generation. Share count has grown roughly 60% in under a year, an ongoing real cost to holders. The Meta compute resale announcement is early stage disconfirming evidence against the assumption that GPU rental demand stays supply constrained. Falsification trigger: if AI Cloud quarterly revenue does not clearly exceed bitcoin mining revenue by the quarter ending September 30, 2026, if additional dilutive capital is needed beyond the May 2026 notes before the Microsoft and Nvidia deployments are substantially online, or if Meta's compute resale materially undercuts pricing for comparable capacity, the bullish case should be considered falsified.
** Bear ($16): delays, a widening financing gap, heavier dilution, bitcoin weakness, and Meta-driven pricing pressure. Base ($40): contracted backlog executes roughly on schedule. Bull ($65): on time execution plus additional large hyperscaler contracts, with Meta's entry proving less disruptive than feared.
Mechanical model output, a screen, not a fair value.conservative screen, understates growth/AI-capex names
Weighing the market price against a scenario range built from the contracted AI hosting backlog (about $3.1 billion under contract from Microsoft, Nvidia, and a Prince George facility, targeting $3.7 billion of annualized revenue by the end of 2026) against the capital, dilution, and now a large unfunded financing gap needed to build it out, the current price sits close to the base case; it is not clearly cheap or clearly expensive.
Research flagged that one or more mechanical-screen inputs look invalid or distorted (e.g. a data-extraction issue or a capex/FCF definition mismatch); the research verdict was independently re-derived from primary filings, not just a check on the same numbers.
Confidence = 0.7 × research sub-score + 0.3 × mechanical sub-score. Mechanical sub-score rewards the Layer 1 screen for agreeing with Layer 2's direction (corroboration), not just for being conviction=NORMAL in isolation.