In the spring of 2025, a solar-powered Airbus subsidiary called AALTO kept its Zephyr aircraft in the stratosphere for 67 straight days, the longest continuous flight ever recorded. It charged its panels in the sun by day and ran through the thin night air on battery power. The cells that carried the second half of that equation were Amprius's.
That single number - 67 days - is the whole thesis in one image. Energy density is the binding constraint on anything with wings in an electric world. A drone, an eVTOL, a high-altitude pseudo-satellite, eventually a plane: their endurance and range are set by watt-hours per kilogram, and mainstream lithium-ion sits around 250–300 Wh/kg. Amprius sells cells rated at 450 Wh/kg commercially and claims an independently validated 500. It gets there by replacing the graphite anode that standard cells use with a silicon anode that can hold roughly ten times the lithium. The more energy per kilo, the longer the flight, and Amprius makes the most.
This is worth taking seriously because aviation does not substitute casually. Qualification runs for months or years, and failure at 70,000 feet is catastrophic - so a cell already flying on the record-setting Zephyr holds an incumbency that a late entrant cannot walk around quickly. The scarce node, in other words, is real.
Now read the economics the way you would read any scarcity claim. A genuine chokepoint gets paid like one: steep gross margins, pricing power over customers with nowhere else to go. Amprius reported a 27% gross margin in its June quarter - a real jump from 9% a year earlier - but still lost $5.1 million in the quarter and booked about a $1 million adjusted-EBITDA loss. Operating margins over the trailing year are deeply negative, and free cash flow is underwater. That is not a toll. That is a scale-up in progress: the structure exists, but the company is not yet collecting rent at the rate its technical lead could justify. The whole investment question sits in whether that 9%-to-27% margin climb continues toward toll territory, or stalls once volume and customer bargaining set in.
The second gap is capacity. If Amprius owned the manufacturing node that gates its own growth, it would be a different story. It mostly does not. Its Fremont pilot line produces on roughly the scale of a demonstration facility, while its real volume flows through contract manufacturers in Asia and its own large-scale Colorado plant is still being built. So its growth today runs through partners' lines - capacity Amprius rents, not capacity only it controls. That, plus a customer list still dominated by a handful of programs, is the difference between owning a scarce capability and merely selling into one.
None of this is to say the demand is imagined. Business is real and moving. In May, Amprius announced roughly $500 million in purchase orders from U.S. defense customers plus a $21 million commercial order. Put that against the more than $140 million in revenue it now guides for all of 2026, and the headline becomes concrete: those new orders alone are about three and a half years' worth of current-year revenue run-rate, a multi-year backlog rather than an annual flow. The quarter before, it named a $24 million order from a European drone developer, cells for Redwire's Stalker Block 30 surveillance drone, and a multi-year electric-motorcycle agreement with Stark Future worth over $100 million. The orders are the easy half. Converting them into revenue, then margin, then cash, before the company needs more capital, is the hard half.
Which brings up price. Ainvest data put the market cap near $1.4 billion on a share price around $9.30 in early October, down hard from a $24.20 fifty-two-week high. That works out to roughly ten times the 2026 revenue forecast for a company still guiding to a net loss. The stock has already given back much of the optimism that ran through 2025, so this is not a case of perfection being fully priced; rather, the remaining multiple assumes the margin ramp and the order backlog keep converting on schedule. Amprius ended the June quarter with about $74.5 million of cash, no net debt, and a big bill coming for its own plant.
Watch the margin line first - it is the cleanest falsifier. If gross margin keeps climbing toward toll territory while revenue holds, the scarce-node story is earning its keep and the derated stock may be the opportunity. If margins stall, orders convert slowly, or the cash runway forces another raise, then the chokepoint was never this vehicle's to collect. The dependency that gave these batteries wings is real and hard to copy. Whether Amprius gets paid for it - and whether the stock you can buy already assumes it does - is the entire argument.