Donut Lab alleged Solid State Battery launched at CES 2026

[Siwastaja]

Wouldn't it be amazing if all the sceptics were wrong and their claims were right. It'll be interesting to see how this plays out.

It would be absolutely wonderful. It's a funny scenario to play with. But don't worry, if it walks like a duck and quacks like a duck... ....

Now a bit more interesting situation would be if they had kind of a good, actual product, not to all those specifications, but still something serious and useful. Let's say, a battery which has 200Wh/kg energy density and can do 10000 cycles with e.g. a bit less ESR drop in cold than the competitors. Now that would be a social shitstorm, with all the skeptics (like me and all other sane people) having to explain "oh, I didn't say that..."

But that's improbable, too. This simply has too many red flags so that most likely it's just a 100% scam with either absolutely no product at all, or in best case, an impractical / useless product of some sort.

The thing is, if you are truly a world-class battery tech developer and able to make such a breakthrough, you would not make it look like a scam just to troll people. You would be enjoying the battery tech, not the trolling. Trolling is usually for losers, so to speak.

[PA0PBZ]

[Siwastaja]

The VTT's report is out:

https://pub-fee113bb711e441db5c353d2d31abbb3.r2.dev/VTT_CR_00092_26.pdf

And it is so far just as I expected: the cell weight is missing, so you can't confirm the key claim of energy density. Apparently, the customer (Donut Lab) forgot to ask the cell to be weighed. How unfortunate.

[ftg]

Heh, exactly.
And now there is a new countdown.
They are going to milk this thing for all it's worth over many weeks it seems.
Or until the first Verge motorcycles ship.

But yep, no mention of cell weight or dimensions. 

[artag]

They are going to milk this thing for all it's worth over many weeks it seems.
Or until the first Verge motorcycles ship.

But surely you've fallen for their evil plan

If it's all hype, they've succeeded in getting you to discuss it and therby promote it. If it's not, the sceptics will end up looking silly.
The right thing to do is to ignore it, expressing no opinion until it's released. Then you can discuss it from a position of knowledge.

Speculation is futile !

[ftg]

But speculation is great fun!
In addition to that, it's a domestic company, so it's a guaranteed coffee table discussion.
And I was not the only one at the office with the countdown on their second or third display today. 
On IRC the speculation is also inescapable.

[Randy222]

Independent testing

https://www.theverge.com/transportation/882993/donut-labs-solid-state-battery-charge-speed-vtt-test

Wouldn't it be amazing if all the sceptics were wrong and their claims were right. It'll be interesting to see how this plays out.

Perhaps if the new tech would somewhat work, and, not be called a battery.
A 4day iphone17 "psu" that can charge in under a minute, perhaps is game changing technology.

[sleemanj]

The VTT's report is out:

https://pub-fee113bb711e441db5c353d2d31abbb3.r2.dev/VTT_CR_00092_26.pdf

And it is so far just as I expected: the cell weight is missing, so you can't confirm the key claim of energy density. Apparently, the customer (Donut Lab) forgot to ask the cell to be weighed. How unfortunate.

This testing was only about charge performance.  They have been plain since starting this "disclosure" that they are doing it piece-meal week by week, obvious for hype reasons, it's a bit annoying but whatever, that's marketing departments for you, always annoying.

Next week they will release another report on another specification, and so on.  Eventually the density will be a weekly focus I expect.

With that said, estimating size of the cell and having a guess at it's weight should be possible from the document and video, actually of course somebody has already taken a stab volumetrically using the cabinet model as a reference https://old.reddit.com/r/DonutLab/comments/1rcm9u4/estimation_of_volumetric_energy_density_from/

[Siwastaja]

With that said, estimating size of the cell and having a guess at it's weight should be possible from the document and video.

Not really to the accuracy needed, though. 400Wh/kg number is exactly what makes all these claims so outragerously significant. At 250Wh/kg, it would be just... meh.

But, a product performing similarly can already be bought off-the-shelf, if 200Wh/kg is acceptable. Without any real reference for measurements (not even a banana), volumetric energy density could be estimated to maybe 50% accuracy if even that, and gravimetric energy density not much at all.

From images you can of course estimate that it's not 20 or 50 Wh/kg or something like that. It's better than that, for sure.

Note that the round-trip energy efficiency at 11C is 82% (and 87% at 5C), which means it does not have miraculously low internal resistance. So if this was real, it would be most usable at 5C max and even then with thermal management, but of course even that is a remarkable number.

But the charge curve, the amount of internal resistance, all mean their sample could be just a state-of-the-art, COTS NCA/NMC high-power cell, overdriven to 4.3V and 11C rating, because in such test, they don't have to care about aging (because it's tested only for a few cycles, not 100000). Heck, 10 year ago I tested COTS LCO and NCA cells charged to 4.45V instead of rated 4.20V, seeing +20% capacity, and no damage / no capacity degradation after 10 cycles, after which I stopped. I could have had VTT to rerun the same test and claim that I have a new battery tech with 20% better capacity, and higher 4.45V voltage, definitely not based on lithium and cobalt, and capable of 100000 cycles. This kind of stunt is easy enough to arrange.

Then again, I'm not 100% sure if even the best COTS NCA/NMC cells behave exactly like this when abused at 11C. I have not personally tested. But they are pretty close, so a real possibility that's all that's going on.

[EEVblog]

The VTT's report is out:
https://pub-fee113bb711e441db5c353d2d31abbb3.r2.dev/VTT_CR_00092_26.pdf

VTT's quote: "which the customer identified as solid-state battery cells"

Interesting to note that they are using voltages that only match Lithium Ion charge profiles, yet, IIRC, they have said that it's capable of varying voltage ranges.

It's almost as if they want you think think it's a solid state battery, but is actually a super capacitor 

Again, these testing houses do exactly what you tell them to, nothing more more, nothing less.

[ejeffrey]

Then again, I'm not 100% sure if even the best COTS NCA/NMC cells behave exactly like this when abused at 11C.

Especially if you can clamp it between two large heat sinks and only have to do it twice.

[EEVblog]

From the report PDF, it was created by the VTT marketing manager on the 9/2/26, so doesn't look to have been tampered with by Donut Labs

[EEVblog]

With that said, estimating size of the cell and having a guess at it's weight should be possible from the document and video, actually of course somebody has already taken a stab volumetrically using the cabinet model as a reference https://old.reddit.com/r/DonutLab/comments/1rcm9u4/estimation_of_volumetric_energy_density_from/

If it was genuinely so good, they would be gloating about the energy density and would have asked VTT to put it in the report and calculate the energy density. That would WOW the world would it not?
So either Donut Labs are dumb and didn't think of doing this, or they are trying to hide it for some reason.

[LogicDude]

I wonder about how much thermal energy was soaked up by the heat sinks.  A big chunk of metal can act as a thermal ground for the battery cell and soak up a lot of heat.

The amount of heat stored in the heat sinks could give you some clues about the charging efficiency so that you could compare it to what was in the report to see if the numbers make sense.

Very few thermocouples were used which I find suspect.  And the copper bars supplying current to the battery cell were another heat sink themselves.

Anyway, it was not a thermal test per se, perhaps that will be done in a future video.

I am not convinced by what we have seen so far.

[thm_w]

Heck, 10 year ago I tested COTS LCO and NCA cells charged to 4.45V instead of rated 4.20V, seeing +20% capacity, and no damage / no capacity degradation after 10 cycles, after which I stopped. I could have had VTT to rerun the same test and claim that I have a new battery tech with 20% better capacity, and higher 4.45V voltage, definitely not based on lithium and cobalt, and capable of 100000 cycles. This kind of stunt is easy enough to arrange.

I don't think many would care as we have 4.35V batteries on the market for 10+ years, and 4.45V for 3+ years. Which gets you more like ~6% capacity increase over 4.35. https://higherwire.com/blogs/articles-1/capacity-difference-between-4-35-and-4-20-volt-cells But I understand what you are getting at.

Hopefully the next test will include cell weight/volume. We already estimated capacity to be ~100Wh so we were close there. And I estimated 70x165x10mm (115cm³). Which is well within the capacity of existing cells on the market.

I think its a bit thicker than I initially thought (12mm?), so the volumetric density is not special (~700Wh/L):

I wonder about how much thermal energy was soaked up by the heat sinks.  A big chunk of metal can act as a thermal ground for the battery cell and soak up a lot of heat.

The amount of heat stored in the heat sinks could give you some clues about the charging efficiency so that you could compare it to what was in the report to see if the numbers make sense.

You can just calculate the round trip efficiency directly as Siwastaja has done already. I would trust the VTT numbers are accurate.

[EEVblog]

Anyway, it was not a thermal test per se, perhaps that will be done in a future video.
I am not convinced by what we have seen so far.

Me neither.
As Siwastaja said, it's quite possible to abuse existing tech to get these results.
They only did the few tests specified by Donut and that's it. No repeat test, no characteristic curves over time and temperature.
VTT didn't do any investigation, or made any independent comment.
Smells like classic "Independent Authority Washing", using VTT as a way to shut everyone up from investigating.

If they really had this revolutionary product, it's easy to demonstrate to the world in a rigorous and convincing way. Yet they have chosen to go about it in the sneakiest way possible.

[EEVblog]

Xjet has make a blog post about it:
https://aardvark.co.nz/daily/2026/0224.shtml

[Simmed]

i will beliv
if jeff dahn verifies it

[booscrawl]

I don't understand everyone's surprise. Quantumscape has posted claims for years that their solid state electrolyte cells can take 4C for thousands of cycles at 45 Celsius without melting and still retain 90% capacity. The "default mode" of QS cells during fast charging is the same as the failure mode for classic wet electrolyte cells: lithium plating.

What really doesn't make sense is that Donut is claiming this is "their battery."

Everyone knows the type of plant needed to produce 100 kWh battery packs for cars is the size of several football fields. They don't  ing have any of those.

The whole thing really steams my hams. What do you mean, you don't believe I made twelve burger patties suspiciously identical to the McDonald's Quarter Pounder in my kitchen in the last seven minutes? But I did make them! No, of course you can't come to the kitchen to watch me make them. And no, I won't tell you which butcher I go to.

[Bud]

Xjet has make a blog post about it:
https://aardvark.co.nz/daily/2026/0224.shtml

...the cell temperature reached almost 90 degrees C despite a heatsink being attached

Ah, now I understand why Donut wants to strap a motorcycle to the battery, the motorcycle is a heatsink! 

The question is do you want to drive a motorcycle with its battery at 90C 

[Simmed]

if anybody notices the chart
there is a thermal lag time due to the thermal mass and distance to sensor
which means there is a even higher temperature at the core
maybe 100 ? 110 ?

[Randy222]

I am curious about the heat part of the charging. Is it getting hot because its charging fast? Is that heat (energy) loss linear across charging amps?
If there's heat being lost during charging, how much joules in heat vs how much joules staying inside the device? If I can charge 150kWh in 1min but also waste about the same, 50%eff is not so great. Can boil that down to just amps, 100amps in but 50amps wasted out in heat.

[EEVblog]

...the cell temperature reached almost 90 degrees C despite a heatsink being attached

Ah, now I understand why Donut wants to strap a motorcycle to the battery, the motorcycle is a heatsink! 
The question is do you want to drive a motorcycle with its battery at 90C 

Airflow for the win!

[Siwastaja]

From the report PDF, it was created by the VTT marketing manager on the 9/2/26, so doesn't look to have been tampered with by Donut Labs

Tampering the document would be absolutely stupid. VTT won't otherwise comment, but if the company published a tampered document they probably would. They would get caught quickly. These guys are not stupid, they know exactly what they are doing and how to manage the publicity.

[rteodor]

The VTT's report is out:
https://pub-fee113bb711e441db5c353d2d31abbb3.r2.dev/VTT_CR_00092_26.pdf

Again, these testing houses do exactly what you tell them to, nothing more more, nothing less.

What if the cells blow up, would that be in the report ? In the video there seem to be two different cells tested. The cell at 4:36 looks different than the cell at 5:01.