IREN Raises $3B in 1% Convertible Notes to Scale AI Cloud After Nvidia and Microsoft Deals

The signal here isn’t just that capital showed up—it’s the price of that capital. IREN has closed $3 billion in convertible notes at a 1% coupon, maturing in 2033, to accelerate its AI cloud build-out following deals with Nvidia and Microsoft. For a business born in Bitcoin mining, where cash flows are cyclical and power markets bite, locking in decade-long money at 1% is a strategic win that few asset-heavy operators secure.

Why this matters: in the miner-to-AI pivot, capital structure becomes the edge. AI infrastructure demands upfront spend—GPUs, power, cooling, networking, and land—well before utilization stabilizes. Debt that looks like equity on the upside and like protection on the downside often clears faster and cheaper than straight bonds or secondary equity. A 1% coupon implies investors are underwriting material equity optionality tied to IREN’s AI roadmap and to the credibility halo from Nvidia and Microsoft. That kind of cost of capital can compress payback periods and widen the feasible set of data center builds.

Convertibles also shape incentives. Equity holders gain rapid capacity expansion without immediate cash dilution, but they accept potential future dilution if the stock rerates. Noteholders secure a low cash coupon with upside if IREN executes. Management gets runway—2033 is far enough out to build, optimize, and reprice workloads—while keeping operating leverage manageable in the near term. In a market where power prices swing and GPU supply remains lumpy, that flexibility is worth more than a headline APR suggests.

Technologically, the bridge from Bitcoin mining to AI cloud is not trivial. Hashrate thrives on batchable compute and can absorb curtailment; AI workloads care about latency, reliability, and SLAs. If IREN aligns capacity with Nvidia’s ecosystem and leverages Microsoft-aligned demand channels, it can shorten the sales cycle and raise utilization faster than a cold-start provider. The funding now backs the hard parts—power contracting, substation buildouts, liquid cooling or high-density air designs, and low-latency interconnects—that turn chips into revenue.

There’s a psychological layer to this raise as well. Markets have been rewarding miners that tell a credible AI story, but the bid fades when capex remains theoretical. A fully subscribed $3 billion convertible at 1% reframes IREN’s narrative from “option on AI” to “funded plan,” which can attract customers who prefer counterparties with committed balance sheets. It also nudges competitors: if peers can’t match the coupon or counterparty confidence, their effective cost of AI capacity rises relative to IREN’s.

From a business standpoint, the thesis is straightforward: transform stranded or low-cost power access and operational muscle into higher-yield compute. AI cloud revenue per megawatt can outstrip mining in many scenarios, but only if utilization and pricing hold. The maturity profile to 2033 gives time to navigate GPU cycles, shift between training and inference mixes, and refine pricing as software stacks evolve. The Nvidia and Microsoft relationships are not just logos; they serve as a signaling device that can lower customer diligence friction and, ultimately, the discount rate applied to future cash flows.

Investors should still calibrate risk. Convertibles reduce immediate cash burden but introduce potential dilution if execution goes right. AI demand is strong today, yet workload mix, model efficiency gains, and potential supply gluts can reset margins. Power policy and grid constraints can delay ramps. And the ethical scrutiny on energy use will intensify as miners reallocate megawatts from hashing to AI. Firms that proactively pair expansion with transparent power sourcing, demand-response participation, and community engagement often enjoy smoother permitting and better customer acceptance.

What to watch next: - Capacity translation: announcements on new megawatts energized and GPU racks installed versus dollars raised. - Utilization and pricing: evidence of contracted workloads, especially tied to the Nvidia/Microsoft spheres. - Balance-sheet moves: any hedging or call-spread overlays to manage dilution risk, and how cash interest remains contained near 1%. - Power strategy: long-dated PPAs, grid interconnect milestones, and curtailment economics.

IREN just bought itself time and flexibility at a rare price point. In AI infrastructure, that combination—cheap, long-duration, option-like capital—often separates the builders that compound from those that chase cycles.