Soluna Finishes Kati 1 and Secures Conditional Base Load Position for All 166 MW in ERCOT

Soluna completed energization of the final 14 MW at Project Kati 1, bringing the Texas green data-center site to its full 83 MW capacity. ERCOT also conditionally classified the full 166 MW Kati campus, including the planned Kati 2 AI expansion, supporting additional capacity for AI and Bitcoin-mining workloads. The milestones strengthen Soluna's operating and expansion infrastructure, though the announcement provides no financial guidance or revenue impact.

Analysis

SLNH’s valuation hinge is no longer incremental Bitcoin-mining capacity; it is whether ERCOT power access can be converted into contracted AI/HPC revenue before the company must fund substantial electrical, cooling, and compute infrastructure. Conditional grid classification is not equivalent to a signed power agreement, completed interconnection, customer commitment, or project financing. The immediate equity reaction can be positive given the scarcity premium assigned to ERCOT-ready load, but the next 1-3 months require evidence of capex funding and an anchor tenant to support a durable multiple re-rating.

The competitive read-through favors larger power-rich digital-infrastructure operators such as CORZ, IREN and WULF, which have better access to capital and more credible paths to monetize power capacity through AI hosting. SLNH’s potential advantage is exposure to curtailed renewable generation, but that model can create utilization and basis-risk volatility if power economics change or ERCOT congestion eases. Over 6-18 months, AI demand could make the site materially more valuable; conversely, equity dilution, debt pricing, or a delay in a definitive ERCOT/interconnection milestone would likely overwhelm the capacity narrative for a small-cap issuer.

Consensus may overvalue the headline MW figure relative to the capital intensity needed to turn it into AI revenue. A credible AI deployment requires disclosed customer economics, power price/pass-through terms, cooling design, delivery timing, and financing sources; absent these, the asset should be valued closer to an option on development rather than as operating AI infrastructure. Thesis is falsified positively by a funded contract with a hyperscaler or well-capitalized GPU cloud customer, and negatively by financing that materially expands share count or by a return to mining-led monetization without contracted AI cash flows.

Key Decisions for Investors

• Do not chase SLNH solely on grid-status news; place on a 1-3 month catalyst watch for a definitive interconnection agreement, named AI/HPC customer, project-level financing, and disclosed MW-to-revenue economics. Upgrade only if those disclosures demonstrate limited dilution and a commercially credible deployment schedule.
• For AI-power exposure, favor a relative long CORZ or IREN versus SLNH over the next 3-6 months: larger operators have more financing flexibility and existing digital-infrastructure credibility, while SLNH retains higher execution and liquidity risk. Reassess if SLNH secures a funded customer contract before peers.
• If trading SLNH, size as a venture-style event position rather than a core AI allocation and use any financing announcement as the key risk gate. Exit or reduce on equity issuance at a material discount, a delayed energization/interconnection timetable, or failure to announce a commercial tenant by the next two reporting cycles.