A metre wide, moving at 50 metres a minute, carrying cathode that will later be slit down into what Celina Mikolajczak calls pancakes. That is a coating line.
Something goes wrong and it takes a minute to notice and react. That minute is 50 metres of material to cut out, and the scrap bins are already full.
Mikolajczak, Lyten's chief battery technology officer, uses that arithmetic to explain why scrap rates are one of the best-protected numbers in the industry. They tell you what a company's margins are, and what they are allowed to be.
Her own view of the ceiling is not hedged. At a 10% scrap rate you cannot survive as a going business. That might be where you sit while you are ramping equipment and fighting every kind of problem, but you cannot stay there without very deep pockets. Scrapped material is straight-up money. If there is no revenue behind it, the company is gone in a heartbeat.
The reason it bites hardest in automotive is volume.
A boutique cell at a fat margin can carry a lot of waste. Automotive is the opposite case: millions of cells a day at margins that were never generous to begin with. Once scrap starts mounting, the profit goes first and fastest. In a gigafactory, Mikolajczak says, you can make a ton of scrap so fast that it costs you a million dollars in a day.
What makes this an operating problem rather than an accounting one is that somebody has to decide, live, what to do about it. If a machine is scrapping a lot, you shut the line down. If it is scrapping a little more than normal, do you still shut down, or keep running and eat the loss? That call sits with the factory leader, and it comes round every single day.
You cannot learn a cell factory from papers
Mikolajczak has been doing batteries for 25 years and describes her career as a move from one hard problem to the next, abandoning each one as it stopped being hard.
She started in late 1999 doing battery failure analysis at Exponent, at a point when there was nobody to teach her how to do it, because the industry had only just discovered it needed the discipline. She watched cell formats turn over from small prismatics and 18650 cylindricals to pouch cells, then went to Tesla, which used nothing but cylindricals, and watched cylindricals come back into fashion as the Model S, Model 3 and Model Y ramped.
The credibility that came out of that work is domestic. She has a vintage Model S, a newer Model Y and three Powerwalls in the garage, roughly 200 kilowatt hours charging at her house every day, and she is entirely relaxed about it. She has burned a lot of cells and run a lot of safety testing, which is where the calm comes from.
After Tesla came pack design at Uber for micromobility and aircraft, then engineering at Panasonic's Gigafactory, then a spell at QuantumScape before Lyten. The Panasonic posting is the one she is most insistent about, and her advice is unusually flat: if you are making cells, or want to, you have to go and get that education, and you cannot get it from papers or conferences. You have to be inside the building.
What you learn there is the pressure of a process that does not stop. The factory runs 24-7, so the problems do not stop at 5pm either. You can slow something down, you can take a machine off line to fix it, but the production days you lose are gone permanently, along with the revenue and the margin attached to them. Some of the answer is in the cell design, which has to be fundamentally manufacturable. The rest is equipment that will run continuously for years, and a team built to keep it running.
Live under that for a while, she says, and the world looks a bit different.
The factory in most people's heads is the wrong factory
Two numbers separate cell making from the manufacturing most people have in mind.
Tolerances in cells are measured in microns. Tact times, at scale, are measured in microseconds. Against that, a car plant works in tolerances of fractions of a millimetre and tact times of a minute or 90 seconds, with people climbing onto the vehicle at each station to do the work.
Nobody touches a cell. The machines touch it, at speed, and they have to be exactly right every time. That is true whichever chemistry you have chosen, and it is what makes the whole business hard.
The picture a lot of people carry around instead is the I Love Lucy episode, two women on a line with their hands on the sweets as they go past. A cell factory looks less like that than almost anything.
Where Mikolajczak parts company with the usual conclusion is on what automation does to employment. The response she keeps hearing is that if it is not manual any more, the workers are not needed. Her answer is that equipment pushing hundreds of thousands of cells a day needs specialists: people who know that at this position and that position the fine dust builds up a little every day, and know exactly what to do about it. That is not an entry-level job you fill off the street.
"This is craftsmanship," she says, and she means it as an argument about who gets to do the work. Cell making, on her reading, is a route back to skilled, respected blue-collar employment rather than a reason to expect less of it.
Nickel is a supply problem before it is a political one
Mikolajczak came to Lyten because of the cathode, and her case for it starts with a constraint rather than a performance claim.
Yes, there is a geopolitical argument about Chinese supply chains, and she gives it a sentence. The more fundamental issue, she says, is that there is a limited amount of nickel, getting more of it means doing increasingly questionable things environmentally, and even if you do those things, the answer to whether there is enough to electrify everything is "not a convincing yes."
She has seen this pattern run once already. Early consumer cells were cobalt oxide, a chemistry she is genuinely fond of, and nobody worried much about the material because cobalt was largely a by-product. As the industry scaled, the conflict-mineral and social problems became impossible to ignore, and the escape route was high nickel. A decade later the industry is standing in front of the same question about nickel.
Lithium sulfur reads differently to her because the questions are boring. Where do you get sulfur, lithium and carbon? The answers are straightforward, and they do not create a fresh environmental problem on the way. She is also unconvinced that lithium itself is genuinely scarce: the industry has taken the cheapest sources, the ones where the sun and the wind do half the work, and has barely started on subsurface brines, which she thinks would drop the price substantially and stabilise world supply if someone chose to pursue them properly. Asked why the oil companies have not, she reports being told the margin is not as good as oil, and does not let that pass. "Well, that's a problem, isn't it?"
The other half of her interest is production. The chemistry runs on standard equipment, and she could see it running in a real factory.
Two arguments for sulfur, and the hard part
The first argument is energy density. Lithium sulfur has a roadmap to roughly twice the gravimetric energy density of high nickel chemistries. Lyten is not there, and she says so: the cells it is making today are at parity with good conventional lithium-ion, somewhere in the 200 to 300 watt hour per kilogram band.
The comparison she draws from that is about slope, not position. Five years of work has got the chemistry to a place conventional lithium-ion took 30 years to reach, which suggests a lot of low-hanging fruit still ahead.
The second argument is cost, and it follows from the simplicity of the supply chain. Sulfur is cheap, so the cells should be substantially cheaper than both nickel chemistries and LFP. Long term, she puts lithium sulfur below $50 a kilowatt hour, on real bill of materials and real cost to produce. Not tomorrow, and not this decade, but eventually, and she does not think the other chemistries have a path to that number at all.
She is sceptical about today's headline prices, which she reads as partly subsidised rather than real. That cuts both ways. Cheap cells have people considering electrifying things they would not have considered before, which is good. An artificially low price also sets up a shock when it corrects, and everyone ends up in a bad place.
The hard part of the chemistry is cycle life. She names it directly, says the strides have been strong and expects cells the market will accept, without claiming the problem is behind her.
Against that she sets the environmental arithmetic. The carbon footprint of a lithium sulfur cell is around half that of a conventional lithium-ion cell today, and should fall further as energy density rises. On recycling, the only thing in the cell worth much is the lithium, so the economics may not always justify the effort, though the process is not especially difficult. And a lithium sulfur cell that ends up in landfill is a much smaller problem than one full of nickel and cobalt.
The margin advantage buys something less obvious, too. Scaling any new chemistry means a period of genuinely bad scrap rates, and cost headroom is what lets a company survive its own mistakes long enough to fix them.
Nobody wins the whole market
Mikolajczak has watched enough chemistries come and go to be sceptical of displacement stories. Cell phones ran on nickel metal hydride when she started. Power tools ran on NiCd, and cadmium is another material nobody wants loose in their community.
Her expectation is that lithium sulfur takes a large chunk of the market rather than the whole of it. High nickel keeps what suits it, LFP keeps what suits it, iron-air of the kind Form is building fills another slot, and lithium air remains, in her word, challenging. What decides each case is cost and performance, the usual things.
The segments she wants are the ones where weight is at a premium: heavy trucks, aircraft, satellites, and micromobility, where taking mass out of the pack decides whether someone will carry a scooter up the stairs to their flat. Then the cheap end, the everyday car and the inexpensive three wheeler, which is where she thinks the average cell price actually comes down, because nickel and LFP will not do it.
And she expects demand to keep opening holes. Ships and other watercraft. More stationary storage. The power draw of the AI build-out, which only works on renewables, and renewables only work with batteries attached. Whichever chemistry suits that job will absorb everything it can get, leaving a gap somewhere else for something to fill.
This piece draws on the full conversation, which is available with a complete transcript on the episode page.