Hieu Duong, lead inventor of a process touted by Tesla CEO Elon Musk as a way to sharply cut the carmaker’s battery costs, has left Tesla for a startup that’s pursuing the same advance using a different approach. In a surprising development, AM Batteries, a tiny Chelmsford, Mass., company with no revenue, hired Duong as chief manufacturing officer and Lie Shi, president of battery separator maker Celgard, as CEO. Both took the jobs in July, but the company did not announce the hires and neither changed his social media profile until this week; the moves have gone unreported until now.
As we have reported, AM is an early-stage developer of dry battery electrode, a method of manufacturing batteries without expensive and toxic wet solvents, the standard way to manufacture cathodes and anodes. AM says its dry strategy is simpler and less costly than Tesla’s, and Duong—the inventor of Tesla’s DBE—calls AM’s approach “DBE 2.0.” On Tuesday, AM was named to The Information 50, our list of the most promising tech startups. With Shi’s recruitment, Yan Wang, co-inventor of AM’s DBE process, has stepped down as CEO and been named the company’s president.
AM’s recruitment of Duong and Shi is a coup, substantially raising the company’s stature in a highly competitive race to cut battery costs and bring down the price of EVs at a time sales of the vehicles are flagging. Duong was lead author of the original 2018 paper on and patents for DBE, which he helped to develop at Maxwell Technologies, a San Diego-based company that Tesla acquired in 2019. At Tesla, Duong became director of electrode engineering, but six months ago, Tesla hired Matt Tyler, previously a senior executive at battery developer 24M Technologies, for the newly created role of director of dry electrode development. That led to confusion outside the company as to who was truly driving DBE development. Duong told me this week that Tyler’s arrival had not created tension and played no role in his decision to leave. He and Tyler led different segments of DBE development, Duong said, and reported to different bosses. “I wanted to do the next big thing,” Duong said. Tyler did not respond to a message requesting comment.
Though Maxwell pioneered DBE, Musk put it on the industry radar in 2020: In a live webcast that he called Battery Day, the Tesla CEO laid out a roadmap for cutting the company’s battery costs by 56%, to around $60 per kilowatt-hour, a price at which virtually any EV could be competitive with a similar combustion vehicle. Central to Musk’s plan was DBE, which he said alone would cut costs by about 15%. The presentation sparked others to try to develop the same money-saving innovation, including Volkswagen, Siemens, Japanese battery maker Asahi Kasei and numerous startups. To date, no one—including Tesla—has figured out how to do DBE at large scale and high speed. “Everyone follows Tesla,” Shi said. “When Tesla says they want to go this direction, everyone wants to chase it.”
The Usual Process, and DBE
In the usual process, workers mix a solvent called N-methylpyrrolidone, or NMP, into a powder containing nickel, manganese, cobalt and lithium, turning it into a slurry that is slathered onto copper foil, creating the cathode. Separately, workers coat a graphite powder mixed with water onto aluminum foil, creating the anode. The foils then pass through calender rollers, flattening the slurry, and into an oven for drying. In the oven, the NMP evaporates and is recovered for reuse.
In DBE, the NMP is eliminated and the dry powders are applied directly to the foil substrates. It’s difficult: Some battery hands have compared the process with attempting to make a cake without eggs, milk or any other liquid.
At Tesla, workers mix the dry powders with a binder called polytetrafluoroethylene, or PTFE (more commonly known as Teflon). This doesn’t turn the mixture into a slurry, but makes the metals stick together like a snowball, Duong said. They then squeeze the powders onto the copper and aluminum, and send it into the calender rollers. The result is an electrode film. That’s where Tesla has found trouble: When workers squeeze the cathode powder mixed with PTFE onto the copper, it’s impossible to make a uniform coating—there are always clumps, valleys and hills of powder. The calender rollers apply tremendous pressure on the mixture in order to make it uniformly flat, but in a 2021 call with investors, Musk said the clumps were damaging the rollers. “The thickest part gets all the pressure from the calender, and that’s how you ruin your equipment,” Duong told me.
AM’s approach doesn’t squeeze the powders onto the foil, but employs a sprayer: Workers charge the powder using electricity, creating an electrostatic field that leads the powder to lay flat on the conductive foils, Duong said. “It drives the particle to the foil,” he said. “You attract it in a uniform fashion.” The foil goes through calender rollers, but only light pressure is required, because the process creates no clumps, he said.
All of this is at early-stage development: AM has only shown the process at a small scale in its lab, and must now assemble a reasonably sized sprayer in order to prove that it may be possible to ramp the concept to industrial scale, Duong said. That seems extremely risky, but Duong said he decided early this year to leave Tesla because the stage of DBE development that excited him—research and invention—was over at Tesla, and the work had turned to ramping up manufacturing. AM takes him back to invention, and both he and Shi told me that AM’s approach amounted to advanced DBE. Improved and scaled up, they said, it could turn into a substantial business. “This opportunity is more exciting than trying to do something people know how to do already,” Shi said. “AM really has a chance to shape the landscape of the whole industry.”
Duong said that AM’s approach to DBE would be cheaper than Tesla’s,and would manufacture electrodes faster. “To me, the AM dry electrode process is dry 2.0,” he said. “It’s more elegant and lower cost. We are pushing the paradigm of technology and cost.”
