Amprius Technologies' second quarter settled, in one print, the question every small-cap battery company eventually faces: can the company turn headline revenue growth into actual gross margin? Revenue of $34.0 million more than doubled year over year (+126%, +19% sequentially), and gross margin landed at 27%, a sixteen-point improvement from a year ago and seven points above the prior quarter - a result that did not exist as recently as the first quarter of 2025, when Amprius was still posting negative gross margin. The quarter also produced the company's first non-GAAP Adjusted EBITDA loss under a million dollars ($-1.0M, a 53% improvement) and a GAAP net loss of $-5.1M that was $1.3M narrower than a year ago, with all of the year-over-year improvement coming from operations and a clean $1.9M non-cash warrant fair-value charge sitting on top of a $-3.2M underlying loss. Management raised the full-year revenue outlook to at least $140 million, lifted the gross-margin guide to at least 28%, reiterated positive Adjusted EBITDA of at least $4 million, and brought forward two new commercial anchors - a $24 million SiCore order from a major European drone developer and a multi-year supply agreement with Barcelona-based Stark Future representing a total revenue opportunity exceeding $100 million starting in 2027.
The setup behind the print is more important than the print itself. Amprius spent four years building a silicon-anode platform that ran pilot-scale silicon in a single Fremont, California line while outsourcing volume through a global contract-manufacturing consortium that now exceeds 2.0 GWh of access, anchored by a former-affiliated silicon-material supplier in Nanjing and a newly formalized South Korean manufacturing alliance. The thesis for 2026 was that the cost curve had to fall off a cliff - because SiCore cells had to compete with graphite-anode lithium-ion on price in markets like commercial drones, e-mobility, and defense, where the energy-density premium is large but the cost premium cannot be. Q2's 27% gross margin - five points higher than management's prior full-year guide of 25%, and now guiding higher to 28% - is the first hard evidence that the curve is bending, and the company is now calling for positive full-year Adjusted EBITDA of at least $4 million after running at a $-2.1M Adjusted EBITDA loss a year ago and at a $-1.0M loss this quarter. The story is not "revenue up" anymore; it is "revenue up AND margin expanded, and management is now willing to print a profitability guide."
The risks are real and structural, not peripheral. The warrant modification added a non-cash $-1.9M to GAAP net loss this quarter, will recur in any quarter where the cash-exercised warrants continue to be marked to fair value, and remains a GAAP-vs-non-GAAP wedge investors must read around. Two customers represented 25% of Q2 revenue and three customers represented 39% of H1 revenue - concentration has improved from 52% / 59% a year ago but the customer base is still narrow. Cash burn was $-40.1M of operating cash flow in the first half, of which $-20.9M of operating-lease cash payments (driven by the January 2026 termination of the Brighton, Colorado lease) is non-recurring; the underlying first-half operating burn is therefore closer to the $-19.2M low-twenties run-rate. With $74.5M of cash on hand and no debt, that gives roughly three to four quarters of runway at the current burn before any further capital raise. The stock has corrected meaningfully off its June 2026 high of $22.91, trading recently at $11.96 - about half of where it sat in early June and back near the lows it printed in mid-2025, despite the operational step-up.
The thesis on the second look: this is a single-product company, on a single technology, with a single Fremont pilot line, riding the first commercial year of a contract-manufacturing-anchored business model - and Q2 was the first quarter in which the gross-margin curve, the cost-reduction road map, the customer mix, the cash position, and the operating-expense ratio all bent in the same direction at the same time. The market has already priced most of the long-tail optionality down to a small-cap battery multiple; it has not yet priced the margin trajectory that the print just made visible. The story the headline tells and the story the financials tell are the same story - and that is the first time that has been true for this company.