Browse
Log in / Register
Share
As of Nov. 19, 2026, at least five battery makers are claiming they hold the world record for energy density. Their published numbers range from 260 Wh/kg to 761 Wh/kg. The odd part is that every one of them is technically true.
This is not a story about cheating. It is a story about a measurement that sounds simple but bends easily. A battery doesn't have one energy density. It has dozens, depending on how fast you drain it, how hard you squeeze it, and how many times you've already charged it. New record numbers appear constantly, but most of them come from a lab bench, not a production line.
Wh/kg is the energy a battery stores divided by its weight. One watt-hour means one watt of power delivered for one hour. If a cell stores 400 Wh and weighs one kilogram, its energy density is 400 Wh/kg. Lighter always looks better.
But there is a second figure, Wh/L, which measures energy by volume. A cell can have the same Wh/kg as another and be much bulkier. SES AI's HAPS datasheet makes that clear: its conventional graphite cell sits at 260 Wh/kg and 700 Wh/L, while its silicon product is rated at 500 Wh/kg and 1,300 Wh/L. Not every press release mentions both numbers, and that omission alone can mislead.
The bigger omission is the C-rate.
C-rate is how fast a cell is charged or discharged. A 0.05C discharge takes roughly twenty hours. A 1C discharge empties the cell in one hour. A 10C discharge does it in six minutes. Every battery performs worse when you push it hard, and the penalty is not small.
SES AI publishes its data across multiple rates, which is rare and valuable. Its 100Ah automotive cell lists over 399 Wh/kg at C/20, 390 at C/3, 370 at 1C, and 351 at 3C. One cell, four numbers, a 13% spread. Its 6Ah drone cell drops from 375 Wh/kg at C/20 to 339 at 1C.
So when a company announces 450 Wh/kg, the first question should be: at what C-rate? Usually the number comes from a gentle trickle discharge. That number is real, but it is not the same as a cell doing real work.
Two more variables hide inside every record claim.
Stack pressure matters especially for solid-state and anode-free cells. A result published by Tsinghua University earlier this year showed 604 Wh/kg and 1,027 Wh/L in a fluorinated polymer solid-state cell with no anode. The caveat: it was measured at 0.05C and needed 1 MPa of external clamping pressure, about 145 pounds per square inch, just to hold the interface together. That is hard to deliver in a car pack.
Cycle life is the third variable. A lab cell can deliver 500 Wh/kg for a few cycles and then fade. U.S. Department of Energy research programs have long treated 80% capacity after 1,000 cycles as a rough threshold for electric-vehicle use. None of the new headline cells disclose that number up front. Drone makers can accept shorter life because payload beats longevity, which is why some of these cells will actually ship for drones before cars.
To sort the claims, ignore the announcement order and look at what each number represents.
The highest published number is 761 Wh/kg, a lithium-sulfur pouch cell with sulfurized polyacrylonitrile, reported in Communications Engineering by Kenji Kakiage's group. It is a genuine record. It was measured at 0.05C and the authors themselves describe the design as too lightweight for practical mass production.
The same chemistry is more believable at 441 Wh/kg when cycled at roughly 0.5C, according to benchmarking work in Nature Communications. That suggests 500 Wh/kg is reachable for lithium-sulfur if the rate penalty can be engineered away.
Below that sits the Tsinghua 604 Wh/kg result from September 2025. It is an important paper, with 274 citations and thousands of readers, but it is not a car battery. It is a controlled laboratory demonstration.
SES AI announced this fall that it is reviving its anode-free lithium-metal cells for drones, with deliveries starting in Q4 2026. Its H6AF pouch cell claims 6.3 Ah, 24.3 Wh, and 525 Wh/kg, and is rated for 7C continuous discharge. Drones are the one market willing to pay for that density and tolerate a shorter cycle life.
Amprius is moving even faster. Its investor materials list 515 Wh/kg as the highest energy density ever delivered to a customer, in a 25 Ah silicon cell developed for a high-altitude platform. The company also won a CES Innovation Award for a 520 Wh/kg design. Its newer SiCore cell reaches 500 Wh/kg at 1C, and the crucial detail is that it can be made on conventional lithium-ion equipment already installed in contract manufacturing plants. Amprius has hit 500 before, but only with specialized processes and never at commercial volume. The company's broadly available SiMaxx product, in production since early 2022, is rated at 450 Wh/kg and 1,150 Wh/L.
That gap between 515 and 450 inside one company is the perfect summary of this problem.
Factorial Energy says it has delivered cells to a global automotive OEM at 390 Wh/kg. Its FEST platform has reached 85% manufacturing yield, up from 10% in early production. The company booked its first commercial drone order in July 2026 and expects first revenue in 2027. That is a shipping product, not a slide deck.
SES AI's 100Ah automotive cell sits at 399 Wh/kg and 862 Wh/L at C/10, dropping to 351 Wh/kg at 3C. Publishing the low numbers for a real product is a sign of honesty, not weakness.
Donut Lab belongs in this tier as a warning. The Finnish startup published an independently measured 409 Wh/kg in September 2026, verified by VTT, Finland's national research center, at a 0.1C discharge rate. The measurement is legitimate. But the company missed its own deadline to have a solid-state cell in a production EV by Q1 2026, and its CEO has said the tested cell is not the one customers will receive. A verified lab result and a shipping product are different claims.
This is the density that actually matters for electric vehicles.
WeLion supplies 360 Wh/kg semi-solid cells to NIO in a 150 kWh pack that gives more than 1,000 km of range in the ET7 and ES6. Once the pack is fully assembled, the system number drops to 260 Wh/kg. That is what the vehicle carries. NIO offers the pack as a subscription, not an option, because these cells remain too expensive for every car.
CATL's Qilin condensed battery reaches 350 Wh/kg and, unusually, it is mass-produced. It uses high-nickel cathodes and silicon anodes, and it has already gone into passenger vehicles. BYD's second-generation Blade Battery is at 162 Wh/kg at the system level, a 5% improvement, and it sells in enormous volume.
Lithium iron phosphate cells still power most new EVs in the world. They cluster around 150 to 190 Wh/kg. They are cheap, safe, long-lived, and good enough. The International Energy Agency projected back in 2020 that conventional lithium-ion would top out near 325 Wh/kg at the cell level. Tesla's 4680 cylindrical cell is roughly 246 Wh/kg.
The often-cited 500 Wh/kg target for solid-state batteries is still exactly that: a target. No automaker has announced a vehicle with a confirmed 500 Wh/kg production pack as of this writing.
Three fall 2026 developments reshaped the ladder. Amprius said its SiCore cell can hold 500 Wh/kg at 1C and be made on ordinary lines, which directly challenges the old excuse that high density requires exotic manufacturing. Factorial reported 85% yield on its FEST platform, a sign that high-density cells can be produced without massive waste. And Donut Lab published a verified record-tier measurement while simultaneously admitting the tested cell would not ship to customers.
Together, those three stories show that the conversation is shifting from who has the highest lab number to who can ship the highest number at 1C for more than a few hundred cycles.
A battery record is only meaningful when you know the discharge rate, the stack pressure, the cycle count, and the production status
