Keel Infrastructure Reports Second Quarter 2026 Results

Keel reports $819M of liquidity (including ~$698M unrestricted cash and ~$121M unencumbered Bitcoin) supported by a $458M convertible note raise during the quarter. Q2 revenue fell 50% YoY to $30.4M and the company posted a $64.0M loss from continuing operations ($0.11/sh) with Adjusted EBITDA of -$24M, reflecting the ongoing pivot away from U.S. Bitcoin mining. Offsetting progress includes permitting momentum across three priority sites and agreements advancing 96MW of Sherbrooke capacity, suggesting improved execution visibility despite near-term earnings pressure.

Analysis

This is less a quarter than a financing checkpoint. The equity should still be valued as an option on scarce power rights, permits, and tenant conversion, not on current operating earnings; until contracts are signed, liquidity mainly funds burn and preserves optionality rather than creating durable equity value.

Second-order winners are the infrastructure incumbents and suppliers that can monetize the same AI/power bottleneck faster: EQIX, DLR, and Vertiv stand to benefit if scarce modules, switchgear, and fiber are repriced by tighter capacity. The loser set is the broader cohort of late-stage developers without balance-sheet depth, because capital and customer demand will migrate toward the few sites that clear permitting first.

Near term, the key catalyst is whether negotiations become signed leases with deposits and start dates over the next 1-3 months; absent that, dilution and schedule creep into 2027 remain the real risk. The contrarian miss is that the Bitcoin wind-down removes upside volatility while leaving the cost base intact, so reported fundamentals can look worse before any HPC revenue appears. A reversal requires tenant-backed project finance, not just permitting headlines.

Key Decisions for Investors

• Short KEEL on any post-print strength or conference-call squeeze; use a 1-3 month horizon and cover only if management announces a signed anchor lease with rent commencement and customer-supported capex.
• Pair trade: long EQIX or DLR / short KEEL to express 'cash-flowed incumbent wins, speculative developer loses' over the next 1-3 months; the risk is a sector-wide re-rating if AI power scarcity tightens faster than expected.
• Buy VRT on pullbacks as the cleaner beneficiary of module and power-infrastructure buildout; this is a better way to express the secular thesis than owning KEEL common before contracts are visible.
• Set an event-driven alert on KEEL for a tenant-backed project-finance package or pre-leasing disclosure; that would be the first credible signal to flip from fade to tactical long.