Layer 0 · IREN — IREN Limited
Thesis
The Bitcoin-miner-pivoting-to-AI-cloud story that just signed a 5-year NVDA infrastructure partnership and is targeting $3.7B ARR by end-CY26. Tier B because the Q1 EPS missed materially (-$0.33 vs -$0.24 est), Bitcoin revenue is decaying, and execution risk on the back-end-weighted ARR ramp is real — but the partnership signature is structural and validates the thesis.
What it does + financial print
IREN owns and operates renewable-powered data-centre capacity at Childress (TX), Sweetwater (TX), and West Texas — originally Bitcoin mining campuses being converted to GPU-as-a-service for AI inference and training. The NVDA partnership locks supply preference and reference-architecture credibility. Microsoft Blackwell deployments are now in flight. The model: convert depreciating Bitcoin economics into multi-year ARR contracts at much higher margins.
Q3 FY26 (May 7) printed $145M revenue (-22% QoQ on Bitcoin price decline) with AI cloud revenue ~$34M, doubling QoQ. EPS missed at -$0.33 vs -$0.24 consensus. The $3.7B ARR target by end-CY26 was reaffirmed but is back-end-weighted. Stock is +33.8% in the last month on the NVDA news and +89% TTM.
Bull case
ARR crosses $1B exit run-rate in Q4 FY26 — that's the visibility checkpoint that confirms the back-end-weighted plan is achievable. Two more hyperscaler contracts at >$500M each TCV. Bitcoin stabilises at >$80k so the legacy mining segment doesn't bleed cash. All three things together get you to Tier A.
Gap / bear case
The EPS miss is not a rounding error — it's nearly 40% worse than consensus and reflects D&A drag from rapid capex on GPU clusters. If ARR ramp slips by a quarter, the stock loses its "growth at any price" bid and reverts to a Bitcoin-miner multiple. Single-customer reliance on NVDA reference-design is a strength and a vulnerability.
Trigger to upgrade / downgrade
Upgrade to Tier A on (a) ARR exit run-rate disclosed >$1.5B in Q4 FY26, OR (b) a second hyperscaler contract signed publicly. Downgrade to C if the ARR target gets pushed to CY27.
The trade
- Entry zone: $42-46 on pullback (50-day moving average zone)
- Stop: $36 close (below NVDA-announcement gap)
- Position size: 1% NLV (small-cap concentration risk)
- Catalyst date: Q4 FY26 print August 27 2026; interim NVDA/MSFT pull-through updates
- Conviction: 6/10