In short: The physical landlord of the AI boom — pre-IPO. Founded 2019 as a SoftBank renewable developer; now an integrated data-center + power platform (Data Centers · Standalone Power · Solutions, incl. acquired Studio 151) that "stops below the GPU": secures land and power, delivers white-space shells on 15–20-year triple-net leases priced on yield-on-cost with annual escalators. ~5 GW of power operating/under construction, three AI campuses (Milam 1.2 GW for OpenAI; PORTS-Pike 8 GW, first phases 2028; Cosmos rent from Q4 FY26). Revenue $232M FY24 → $214M FY25 → $139M H1 FY26 (+66%); $3.2B net loss mostly non-cash ($2.6B warrant revaluation, $590M SBC), ~$56M operating cash used. $439B backlog, $357B beyond year 8, ~$178B of CapEx to build it; PORTS-Pike alone needs 9.2 GW of new gas generation with turbine lead times up to seven years. Risks: execution (abatement/termination rights), customer concentration, financing spread, power. Personal take: at a reported >$50B, "I'd want to see the first major campuses operating and the project-level economics proven before getting comfortable with the price."
SB Energy started out building solar farms and batteries for SoftBank. It is now building the giant buildings that house AI computers — and the power supply that feeds them — and renting them to AI companies, mainly OpenAI. It does not buy the chips: the tenant brings those. SB Energy is essentially a landlord, and its leases are "triple-net," meaning the tenant pays the building's running costs while SB Energy collects rent for 15–20 years.
The headline numbers look wild: $139 million of revenue in six months, but $439 billion of signed future rent. The catch is that most of that rent arrives more than eight years from now, none of the big data centers is running yet, and building them will cost about $178 billion. The reported $3.2 billion loss is mostly a paper charge — the warrants (rights to buy shares) it gave OpenAI became more valuable as SB Energy did.
Whether shareholders make money comes down to one spread: the author says developers aim to earn roughly 8–11% a year on what they spend to build, and borrow at 6–7%. That gap is thin enough that cost overruns or pricier loans can erase it. At a reported valuation above $50 billion, the author would wait until the first campuses are actually operating and earning. Analysis, not a recommendation.
Nothing matches this filter.
Verbatim excerpts from the public transcripts (auto-pulled at each mention's timestamp, lightly cleaned). Timestamps link into the video; "source page" opens that commentator's full analysis at this row. Click a mention's header line to expand it (one open at a time). For personal study — not investment advice.