The electrode line in a gigafactory ends in an oven about a hundred metres long, and the only thing that oven does is drive off a solvent.

The solvent is NMP. It is toxic, so the plant needs a recovery system to keep it inside the building, and it boils at 200 degrees centigrade, which is why the oven has to be that long. Battery makers tell AM Batteries that more than 40% of the energy their plants consume goes into drying. Producers that cannot reprocess the recovered NMP sell it on to chemical companies, and buy more.

AM Batteries sells equipment that deletes the step. Rather than mixing active material, conductive additive and binder into a slurry and coating it onto the current collector, its process deposits the dry powder onto the foil electrostatically. No solvent, no slurry, no oven.

The company does not make cells. It sells turnkey equipment to the people who do.

What Tesla's battery day did for a company that was writing papers

AM Batteries was founded in 2016 by two professors, one from Texas A&M and one from Worcester Polytechnic Institute, with a single question: how do you get the solvent out of electrode making. For roughly four years the answer stayed in the lab and in journals.

What changed was Tesla's battery day in 2020, where Elon Musk put dry coating at the centre of the 4680 cell and the manufacturing cost reduction that was supposed to come with it. Lie Shi is direct about what that meant for a small company with a solvent-free process and no customers: the founders realised their moment had arrived. A seed round followed within twelve months, then a Series A, then $30 million in October 2023 led by Toyota Ventures. The company moved into a new facility and began building engineering pilot lines.

He joined about fourteen months before this conversation, after more than twenty years in the industry and a spell as president of Celgard, one of the largest separator producers in the United States. A board member called him. His reason for taking it is competitive rather than technical: you do not get many chances in a career to compete against Tesla in a new field.

Every number is measured against the wet line

Energy consumption, down 40%. Capital expenditure and operating expenditure, down 40% as well. Floor space, down 70%, which he attributes to AM Batteries' particular approach rather than to dry coating in general.

Carried through to a finished cell, he puts the saving at 10 to 15%. On a cell costing $100 per kilowatt hour that is $10 to $15.

His argument for why this lands now is about what is left to cut. The easy savings in cell manufacturing have already been taken, so a process that offers a reduction of that size gets a hearing it would not have got a decade ago.

There is a second gain, and it comes from thickness. Energy density is still set by the cathode and anode, but wet lines rarely coat thick electrodes, because uniformity suffers and thicker coatings cost more to dry. Most sit around 60 microns. Dry deposition can load far more: AM Batteries has run five times the wet standard in trials, and reaches 120 to 130 microns in normal use. It does not need to go to the extreme. Simply doubling the thickness gets 5 to 10% more energy density at cell level.

The one chemistry it cannot do

At the level of physics, Shi is comfortable calling the process chemistry agnostic. The company has run NMC at both 622 and 811, LFP, silicon anode and some sodium-ion materials, and all of them work.

The exception is lithium metal anode. Deposition is electrostatic, and a conductive material will not hold the powder onto the foil.

The harder problem is not chemistry, and he does not pretend otherwise. AM Batteries faces the same scale-up challenge everyone in dry coating faces, and it reduces to three things: making the film wider, making it uniform, and running it fast enough to compete with a wet line.

Days before this conversation the company shipped rolls of film to a customer for the first time, made on the engineering pilot line. Until then it had sent sheet samples. Rolls mean the customer can build a large pouch cell and compare it against wet process electrodes directly, which is the comparison that decides anything.

The pitch to what he calls lighthouse customers, a phrase he says he learned recently for early adopters, is deliberately modest. Keep your binder. Keep the same binder percentage and the same active material loading. Change the machine and nothing else. Battery makers changing one thing do not want to change three, and when they run the comparison, he says, they report equivalent rate capability, energy density and cycle life.

The more ambitious version is next. AM Batteries has signed a joint development agreement with Zeon, a major binder producer, to develop a binder made for a dry line rather than adapted from a wet one. Today's binders carry additives put there to make the wet process work. His instruction to Zeon was to take those out.

Why the process avoids PTFE

He wants one distinction on the record, because the comparison with Tesla is otherwise inevitable.

Tesla's dry process came out of Maxwell, which it bought in 2019 and whose supercapacitor business it sold back into the market two years later, keeping the coating technology. That route depends on PTFE, and PTFE is a difficult material to work with. It is why the mixing step is so energy intensive, and it is not stable in the anode.

AM Batteries does not use it. Whatever binder the customer specifies, the company grinds it to the right particle size and controls its flowability and how it behaves in electrostatic deposition.

The second half of the distinction is commercial. Tesla makes cells and cars; AM Batteries makes equipment. They are not in the same market, and he is willing to say out loud that Tesla could one day be a customer.

Which does not mean he is uncompetitive about it. The company hired, from Tesla last year, the engineering director who invented the dry battery electrode process at Maxwell. First catch up with Tesla commercially, is how he frames the sequence, then leapfrog them.

Five years to a percolation point

His timing estimate, drawn from conversations with other manufacturers, is five years to a percolation point at which the major players all adopt, and ten years to dry coating being general across gigafactories.

Talking to those manufacturers has told him something else: everyone is already working on dry electrodes. The reactions to AM Batteries' own process fall into three groups. Sign a joint development agreement and begin qualification. Go with an equipment supplier that already sells to Tesla, on the grounds that Tesla has proved the approach works. Or develop it in house, which is mostly the Japanese and Korean answer.

He does not expect the wet process to disappear. High precision consumer electronics still need the uniformity it delivers, and uniformity is a safety question as much as a performance one. He reaches for the Galaxy Note 7 as the example: separators taken down to five or seven microns, uniformity not good enough, and the consequences that followed.

Automotive and stationary storage are a different case, and he thinks both should move to dry quickly, for the reason that governs everything else in this conversation. From his automotive customers, it is all about reducing cost.

This piece draws on the full conversation, which is available with a complete transcript on the episode page.