Fermi (FRMI) Q2 2026 Earnings Call Transcript

Fermi signed a binding 15-year anchor lease with TensorWave for 222 MW initial capacity (up to 650 MW total), targeting ~$6.5B of revenue over the initial term, and reported a Q2 net loss of $25.8M ($0.04/share) driven largely by $26.8M in G&A. The company also strengthened liquidity with a $431M (5% coupon) convertible senior notes offering due 2031, generating ~$382M net proceeds, while three Siemens Energy F-class turbines (up to 728 MW each in simple cycle) arrived at the Port of Houston to support first power of 210 MW by July 1, 2027 and 640 MW nameplate by Q4 2027. Management highlighted a Hillcore “power alliance” adding ~2.6 GW (doubling total site power to 4.8 GW) and stated variable energy charges are fully passed through to customers, reducing commodity-fuel risk.

Analysis

This is more important as a proof-of-concept for private-power AI campuses than as an immediate cash-flow event. The real read-through is that capital will likely crowd toward the bottleneck layers of the stack — long-lead turbines, EPC, gas infrastructure, and project finance — while grid-dependent colocation models lose relative scarcity value. For FRMI equity, the upside is from re-rating a scarce asset, but the economic leak is that most of the announced value still sits behind execution gates rather than in today’s earnings power.

Near term, the stock is trading on financing optics, not operating cash generation. The next 30-60 days matter more than the next 12 months because the backstop and project-finance milestones are the true de-risking events; if either slips, the market will quickly reprice this as a highly levered development story again. Over 6-18 months, the key variable is whether first power lands on time — even a modest schedule delay would likely force incremental dilution or debt, compressing the multiple and shifting value to lenders/suppliers.

The contrarian view is that the market may be over-assigning certainty to a contract that is economically real but operationally contingent. The capped-call structure only softens dilution if the shares rally hard; it does not eliminate the financing overhang while construction is still ahead. XEL is only a marginal beneficiary from the utility tie, so the cleaner expression is not a utility long; it is a disciplined wait-for-confirmation posture on FRMI, or a fade if the stock gaps on headlines before the financing stack is fully closed.

Key Decisions for Investors

• Do not chase FRMI on the headline; wait for the backstop and project-finance announcements before initiating risk. If those clear within 30-60 days, use a 3-6 month call spread to express upside with defined downside; if they slip, the thesis is impaired.
• If already long FRMI, trim into post-event strength and keep only a core tracking position until the 2027 first-power schedule is independently de-risked. Falsifier: any delay to the first-power target or evidence of incremental equity need.
• Avoid using XEL as the main expression; the utility linkage is too small to matter to earnings. Treat XEL only as a monitoring vehicle for whether regulators are broadly tolerant of adjacent load growth.