On 2 September 2026 a Taiwanese company said the sentence the battery industry has been rehearsing for a decade. ProLogium Technology announced that its Generation 3.5 Lithium Ceramic Battery had entered mass production at its plant in Taoyuan, which would make it an all-solid-state cell coming off a line rather than a pilot bench.
The headline numbers are unusually well documented. A TUV test report puts the 185.4 Ah large-format cell at 381 watt-hours per kilogram and 903 watt-hours per litre. UL Solutions separately ran the same cell through China GB/T 43568-2026, the national method written to decide what may legally be called solid-state.
What deserves a closer look is not the energy density on its own. It is that the claim arrived with paperwork from two outside laboratories, and that the factory standing behind it produces half a gigawatt-hour a year. Both halves of that sentence matter, and they pull in opposite directions.
Solid-state battery mass production, what the two test reports measured
TUV Rheinland handled the energy density. The cell it measured is a 185.4 Ah pouch, which is the size range carmakers actually specify, not a coin-sized research sample that flatters a spreadsheet.
UL Solutions handled the harder question. Under GB/T 43568-2026 the cell sat for six hours in a continuous vacuum at 120 degrees Celsius, and ProLogium reports it lost under 0.05 per cent of its mass against a 0.5 per cent ceiling.
The logic of that test is simple. Liquid electrolyte evaporates under heat and vacuum, so a cell that barely loses weight cannot be hiding much liquid inside. It is a crude instrument, but it is a measurable one.
That matters because the word solid-state had been drifting. Chinese regulators wrote the standard in part because packs kept arriving with the label attached while still carrying liquid electrolyte, most visibly when the SAIC-backed brand IM Motors promoted a QingTao-developed pack as a solid-state product.
China has since submitted GB/T 43568-2026 to the International Electrotechnical Commission as a reference for international standardisation. Whatever one thinks of the origin, the industry is about to get a shared line in the sand.
A decade of small lines is the actual news
ProLogium settled its cell architecture in 2012 and has not moved off it since. The design, which it calls Logithium, pairs a ceramic separator with an edge frame around the electrode that provides sealing and insulation and helps isolate burrs.
Commercial production started in 2013 on a sheet-by-sheet line aimed at wearables, medical devices and semiconductor equipment. Those are small, unglamorous markets, but they pay, and they teach a company about yield.
A roll-to-roll line followed in 2017, and the giga-level platform entered operation in 2024. The company says cumulative shipments across the platform now exceed 2.4 million cells.
That figure needs unpacking rather than repeating. The bulk of the disclosed automotive business is a US car-audio supplier whose parts go into vehicles built in North America by a Japanese carmaker, with more than 175 repeat orders and over 900,000 cells delivered.
Those are not traction cells for a high-voltage pack. What the record does prove is that the company has shipped to automotive quality expectations for years, holding IATF 16949 certification since 2022 and passing annual audits.
The next generation is where the platform argument gets interesting. ProLogium says its Gen 4 cell, built on a fully inorganic electrolyte system, keeps the same architecture and would need roughly 10 per cent of the existing line modified to build.
Gen 4 is also where the safety story lives. The company describes an active mechanism that stabilises the electrode materials at high temperature to head off thermal runaway, alongside better cold-weather behaviour and lower material cost.
The cost wall between a good cell and a cheaper car
Half a gigawatt-hour a year is roughly 6,250 packs of 80 kilowatt-hours, if every cell went to cars, which it will not. The company plans to grow the Taoyuan site toward one to two gigawatt-hours.
The real scale sits in France. A plant at Dunkirk is designed for a four gigawatt-hour first phase ramping toward 2030, with a maximum design capacity of 44 gigawatt-hours, which is eight times the current Taiwanese line in its opening phase alone.
North America gets a lighter approach. ProLogium is evaluating shipping inlays, the single-layer unit at the core of the cell, from France to partners who would stack them into pouch cells and build packs closer to customers.
Then there is price. Mass-market lithium iron phosphate cells in China have been reported around 300 to 400 yuan per kilowatt-hour, roughly 42 to 55 dollars, and ProLogium has not published a comparable figure for Gen 3.5.
Pack engineering will take a further bite. Cooling, structure, electrical protection and management hardware all add mass and volume, so a striking cell number never survives intact to the vehicle.
The competition is not standing still either. Chery has said its solid-liquid hybrid cell enters vehicles late in 2026 with all-solid-state validation in 2027, while BYD, CATL and Geely are aiming at trial production of solid-state cells in 2027 along a sulfide-based route.
Read across the whole field and the bottleneck everyone now names is the same one: not chemistry, but the line that has to build the chemistry a million times without variation. That is the axis Tesla picked years ago, when it spent almost a decade on dry electrode processing for the 4680 cell instead of chasing a headline chemistry.
Giga Texas now turns out well over a million cells a week, and Tesla has locked domestic LFP supply for its storage products rather than waiting for a laboratory breakthrough to rescue the cost curve. The same instinct shows up in how it approaches low-cost electric vehicles, where the factory is designed before the product is shown.
ProLogium has arrived at that conclusion from the other direction, and its press release almost says so out loud: the point of keeping one architecture across four chemistry generations is that industrial capacity never has to be rebuilt. On that measure, 381 watt-hours per kilogram is a real result and an unfinished one.
References
ProLogium Technology, press release, 2 September 2026 · CarNewsChina, Adrian Leung, 6 September 2026 · Battery-Tech Network weekly briefing, 6 September 2026 · Interesting Engineering, 2 September 2026 · ProLogium board and financing announcements, 28 July and 24 August 2026
