Keel Infrastructure (KEEL) Q1 2026 Earnings Call Transcript

Keel Infrastructure reported Q1 revenue of $37 million, down 23% year over year, with operating loss widening to $98 million and adjusted EBITDA turning negative $17 million. Offsetting the weaker earnings, the company said it has about $533 million in cash and Bitcoin, enough to fund Panther Creek, Sharon, and Moses Lake through lease execution and start construction at Moses Lake while covering G&A through 2028. Management reiterated its goal to sign three leases by year-end and highlighted progress on permitting, site development, and the ongoing wind-down of Bitcoin mining.

Analysis

The market is likely underappreciating how much of this story is now about option value on power, not current earnings. The balance sheet gives them a long runway, which matters because in scarce-power markets the winner is often the developer that can wait out permitting and still show up when tenants are ready to sign. That creates a second-order advantage versus smaller peers that may have power but not the capital to hold land, equipment, and permitting optionality through 2027.

The key catalyst is not “growth” in the abstract; it is lease conversion, and the asymmetry is large. Once a single credible lease lands, the financing mix can shift from corporate resource drain to project-level, nonrecourse capital, which should compress the equity discount quickly. The flip side is that every quarter without a signed lease raises the risk that the market re-rates this as a perpetual pre-revenue developer rather than a monetizing infrastructure platform.

The strongest hidden beneficiary here may be PPL: secured load in a tight PJM environment creates a multi-year demand anchor and improves the strategic value of local grid assets, substations, and interconnect capacity. On the semiconductor side, the explicit design around next-gen GPU infrastructure is a subtle positive for NVDA, but more as a capacity-enabler than a direct revenue driver. The more important implication is that power-constrained HPC demand could stay structurally elevated longer than consensus expects, because supply is being rationed by delivery timelines rather than by end-demand.

Contrarian risk: the stock can still work even if the business remains ugly for several more quarters, but only if management keeps converting narrative into contract. The biggest failure mode is not lower demand; it is execution slippage at permitting, tenant mix, or capital deployment, which would leave cash burning faster than lease monetization. That makes this a time-sensitive situation over the next 2-3 quarters, with the real inflection window into year-end lease announcements and 2027 project financing.

Key Decisions for Investors

• Long KEEL common into lease-announcement catalysts over the next 2-3 quarters; target a re-rating if even one site signs, but size modestly because the path remains execution-driven and binary.
• Use call spreads in KEEL rather than stock if available: structure 6-12 month upside exposure to the year-end lease window while defining downside if permitting or tenant selection slips.
• Pair trade: long PPL / short a basket of legacy crypto-miners or stranded-power names over 3-6 months; this thesis favors regulated utility load monetization over melting-icecube mining economics.
• Long NVDA on any weakness tied to AI power-scarcity headlines; the second-order effect is more demand for advanced GPU-ready infrastructure, which supports system-level adoption even if KEEL itself remains pre-revenue.
• Avoid chasing the equity if management misses the mid-late summer permitting window; the setup degrades quickly if the market starts discounting 2027 cash flows back into 2028-29.