What Pelosi Just Bought in AI Power

Congress just put a familiar name on one of the hardest fights in the market right now: who actually powers the AI boom when the grid cannot keep up. A new Periodic Transaction Report tied to Rep. Nancy Pelosi shows her household bought Bloom Energy (BE) stock and long-dated call options in late July. It is the first disclosed Bloom position for that household. The paperwork landed this week, and traders treated it like fresh oxygen on a name that had already gone parabolic earlier in the year and then cooled off hard from the highs. I do not invest because a politician buys something. I do pay attention when a high-profile buyer shows up in a story that was already real before the filing hit. ## The real problem Bloom is selling into AI data centers do not just need more chips. They need more power, faster, with less drama about interconnects, permits, and how long a gas turbine order sits in line. Bloom builds solid-oxide fuel cell systems that make electricity on site. Think factory-style boxes that stack, ship, and plug into critical load instead of waiting years for the perfect grid upgrade. Management has been loud about the same customer pattern for months: operators who once defaulted to combustion gear are switching because fuel cells can get installed and approved faster in a lot of real projects. That is the kitchen-table version. Data centers are starving for reliable megawatts. Bloom is trying to sell time-to-power, not a science-fair hydrogen dream. ## What the July numbers already said Before anybody debated Pelosi's disclosure, Bloom's second quarter already forced a second look. Revenue hit about $1.065 billion. That was the company's first billion-dollar quarter, up roughly 166% from a year earlier. Product revenue did most of the heavy lifting, climbing about 215% as AI and large commercial power customers converted orders into shipments. Non-GAAP earnings came in around $0.78 a share, well ahead of what Wall Street had been modeling heading into the report. Management raised full-year 2026 revenue guidance to $3.9 billion to $4.2 billion. At the midpoint, that is roughly a double versus last year. They also pointed to a customer base that now includes the major U.S. hyperscalers plus a growing list of neoclouds, AI labs, and colocation operators who have validated the gear for AI factories. There is more color if you read past the headline. Bloom has talked about a multi-gigawatt master agreement path with Oracle, a sharply larger Brookfield financing partnership for onsite data-center power, and internal visibility into a much bigger multi-year deployment pipeline than the company could talk about two years ago. None of that makes the stock cheap or safe. It does explain why the name stopped trading like a forgotten fuel-cell science project. ## Why the filing still matters Pelosi's report covers purchases on July 24 and July 28: 15,000 shares of Bloom common stock plus 200 call options struck at $100 and expiring in mid-June 2027. Disclosure ranges put the household's Bloom exposure in the multi-million-dollar band, not a token tip-toe. The same filing also showed fresh Intel activity, which is a separate chip-manufacturing story. The timing is the part people will argue about for weeks. Late July sat near a softer stretch for the stock after a huge run. The disclosure itself arrived in late August, which is normal for this kind of paperwork and still late enough that copycats are buying the headline, not the original fill. When Wall Street loves a name, I get skeptical. When it hates one after a monster rally, I get skeptical too. The Street gets paid on volume and fees. A famous buyer does not fix dilution, execution risk, or a valuation that already priced a lot of perfection near the June highs above $350. Shares recently traded near the low $200s, still miles above last year's levels and still a long way under the peak. So strip the politics out for a second. The filing is a spotlight. The business question underneath it is simpler: are AI builders still choosing onsite fuel cells because the grid is too slow, or did the easy part of the trade already happen? ## What can still go wrong A lot. Bloom is no longer a tiny story stock, but it is still a company that has to manufacture, install, and service hardware at scale while customers argue about fuel supply, site design, and who carries project risk. Guidance this high leaves less room for a shipping miss. Competition from turbines, engines, batteries, and other onsite power vendors does not disappear because one quarter printed huge. The stock already taught that lesson once this summer. A name can grow revenue at triple-digit rates and still get cut in half from the top when crowded money decides the next ten years of AI power got priced into one quarter. Short-seller attacks, lawsuit headlines, and valuation fights all travel with a stock that has already made people rich on the way up. I am also no Washington accountant. Congressional disclosures are lagging by design. Copying a famous name after the fact is a great way to buy someone else's leftover excitement. ## Bottom Line My take: respect the AI power demand story Bloom is actually selling, treat the Pelosi disclosure as a late spotlight instead of a free pass, and remember how hard this stock already punished anyone who bought the peak. Bloom's customers are trying to solve a real American infrastructure problem: getting enough clean, fast power to the racks without waiting forever on the grid. The second-quarter numbers show that pitch is landing with serious buyers. Whether that still moves the needle from here depends less on the next political headline and more on whether Bloom can keep shipping at the pace the guidance now assumes.