IREN Limited's Sweetwater Hub: Can It Fuel IREN's AI Expansion Plans?

IREN plans about 2 GW of gross power capacity at its Sweetwater campus: 1.4 GW at Sweetwater 1 and 600 MW at Sweetwater 2, using a liquid-cooled design it aims to replicate across future projects. IREN says recent three-year AI Cloud contracts generate more than $20 million per IT MW, with discussions around $25 million, but those figures are not Sweetwater-specific; execution depends on capital, ERCOT grid access and converting planned capacity into revenue. Contractual commitments totaled $13.81 billion as of June 30, 2026, while fiscal 2027 earnings estimates were revised down in the past week; shares gained 12.1% over six months and trade at 3.86x forward sales versus a 2.70x industry average.

Analysis

The key valuation risk is converting gross site power into billable IT capacity: losses from cooling and power delivery, plus GPU availability and commissioning delays, can make headline campus MW a poor proxy for near-term revenue. Standardized liquid cooling could lower repeat-build friction, but only after Sweetwater 1 demonstrates on-time commissioning and usable capacity; until then, replication is an execution hypothesis, not proven operating leverage.

IREN’s premium to the industry alongside recent FY2027 estimate cuts leaves less room for schedule slippage. The cited contract economics are not Sweetwater-specific, and the $13.81B commitment figure needs diligence on what it represents—contracted spending versus financing or other obligations, timing, cancellation terms, and customer concentration—before treating it as either secured demand or balance-sheet debt. Power interconnection, transformers, construction capacity, and GPU allocation are potential bottlenecks; earlier delivery by CoreWeave or Nebius could capture customers even if IREN’s eventual design is competitive. Conversely, peers also face power and capital constraints, so their expansion plans are not proof of executable capacity.

Near term (days), the estimate revisions and premium create downside sensitivity to further cuts; the article itself offers no new operating milestone to justify chasing the shares. Over 1–3 months, watch financing terms, firm ERCOT interconnection milestones, and disclosed customer/IT-capacity conversion. Over 6–18 months, commissioning pace and utilization determine whether the standardized design earns a repeatability premium. The contrarian upside is that markets may discount the value of a repeatable liquid-cooled platform—but it should be credited only after measured delivery. The thesis weakens materially with funded construction, dated grid milestones, commissioned IT MW, and estimates stabilizing; further estimate cuts or missed milestones falsify the execution case.

Key Decisions for Investors

• Do not chase IREN on planned gross capacity alone. For existing exposure, keep sizing disciplined; consider trimming into strength while estimates are falling and valuation remains at a reported premium, rather than initiating an outright short without fresh evidence of slippage.
• Set an add-back alert—not a price target—on independently verifiable milestones: funded buildout, dated grid connection, commissioned IT MW, and customer contracts tied specifically to Sweetwater. Confirm whether disclosed capacity is gross site power or revenue-capable IT load.
• Track IREN’s next disclosures for the composition and timing of the $13.81B in commitments, cancellation provisions, customer concentration, and capital sources. Escalate downside risk if commitments require materially more funding than expected or FY2027 estimates are cut again.
• Keep CRWV and NBIS as relative-execution comparators, not automatic longs: compare delivered capacity, utilization, customer economics, and funding terms before expressing a pair trade.