IREN thesis brief

• IREN is now primarily an AI data-centre / power / GPU-cloud thesis, not a bitcoin-miner thesis. BTC cashflow matters, but it is no longer the main reason to care.
• The Microsoft contract changed the shape of the company: $9.7B over five years, 200MW, GB300 infrastructure, four phases through 2026. That is a serious demand anchor if execution holds.
• The Dell GPU purchase and NVIDIA-related signals support the idea that IREN is moving into credible AI compute supply, not just rebranding mining facilities.
• The hard part is financing. Convertibles, capped calls and dilution are turbulence if they fund contracted growth; they become thesis damage if terms worsen or deployment slips.
• Patrick’s May 22 reframe stands: this may be the closest current analogue to early RKLB — volatile, easy to over-manage, but potentially worth accumulating while the market argues about the old business.

Watch next

• Horizon phase deployments and whether Microsoft capacity comes online on schedule.
• Sweetwater/Oklahoma power-campus progress and interconnect/permitting signals.
• Financing terms: coupon, conversion premium, capped calls, share count impact.
• AI ARR / contracted revenue conversion versus company targets.

Main risks

• Execution/capex risk is huge; this is infrastructure buildout under pressure.
• Repeated financing can dilute upside if growth is not fast enough.
• BTC weakness can still affect cashflow and sentiment even if thesis has pivoted.
• Customer concentration if Microsoft dominates the story.