Back to News
Market Impact: 0.56

Why IREN Limited Rallied Almost 40% in May

Artificial IntelligenceCorporate EarningsCompany FundamentalsM&A & RestructuringTechnology & InnovationMarket Technicals & Flows
Why IREN Limited Rallied Almost 40% in May

IREN reported fiscal Q3 revenue of $144.8 million, down 21.6%, while adjusted EBITDA fell 20% quarter-over-quarter to $59.5 million, but the company simultaneously advanced its pivot to AI cloud infrastructure. May included two acquisitions—Mirantis for $625 million in stock and Nostrum Group for an estimated 165 million euros—as well as a Nvidia partnership covering up to 5 GW of DSX deployment, a $3.4 billion cloud contract, and a warrant for 30 million shares at $70. IREN also raised nearly $3 billion in 1% convertible notes, helping fund its build-out despite ongoing execution and capital needs.

Analysis

The market is starting to price IREN less as a marginal miner and more as a scarcity asset: a large, power-secured land bank that is increasingly relevant to GPU deployment. That re-rating matters because the bottleneck in AI infrastructure is shifting from silicon availability to energized capacity, permitting, and time-to-connect; any credible endorsement from Nvidia effectively compresses perceived execution risk and raises the odds of strategic optionality, including a takeout premium. The trade is therefore less about current earnings power and more about who controls future scarce megawatts in an AI-constrained market.

The second-order effect is that IREN’s customer mix may become less important than its financing stack. The business can now plausibly access cheaper capital than a standalone convert issuer, but dilution risk is still not eliminated — it is simply deferred until milestone credibility improves. The recent financing structure suggests equity upside is being used as a currency to buy time; if buildout slips even modestly, the market will likely punish the name harder than before because expectations have shifted from turnaround to platform transformation.

For Nvidia, the partnership is strategically useful even if the direct economics are modest: it extends control deeper into deployment, potentially bypasses hyperscaler rent extraction, and creates a preferred venue for custom AI factory architecture. That said, if the market starts treating the relationship as quasi-captive supply, the bear case becomes that IREN’s economics could be squeezed by a dominant counterparty while still carrying heavy capex obligations. The clearest catalyst path is over the next 3-12 months: additional power activation, contract conversion, and financing at progressively lower spreads; the key reversal risk is any delay in energization or a cooling in AI infrastructure demand before the next tranche of capacity is monetized.