Inside the Small Factory Where General Motors Is Building the Future of Its Electric Vehicles

If you are driving across the General Motors campus in Warren, Michigan, you will likely overlook two anonymous white buildings next to each other, located in the upper part of the 710-hectare site. Inside these buildings, which total half a million square feet, a massive construction project is still underway. Upon completion, the location of […]

If you drive by General Motors' sprawling campus in Warren, Michigan, you're likely to miss two adjacent, nondescript white buildings perched atop the 710-acre site.

Inside these buildings, which total half a million square feet, there is a massive construction project still underway. Once completed, it will house General Motors' new battery cell development center, where the automaker is aggressively building new prototype cells aimed at lowering costs for its electric vehicles.

GM needs to build faster and cheaper to achieve its goal of selling affordable electric cars in much higher volumes. These EVs will have to compete not only with longtime sales leader Tesla, but also with the flood of Chinese EVs that will sooner or later land in the US. And while electric vehicle sales in the U.S. appear to be slowing, America has a long way to go before it can reliably compete with China, the world's leading EV market.

With gas prices rising 40 percent this year as a result of the war in Iran, there has been a surge in buyers interested in new and used electric cars — but mostly in models priced below the $50,000 average price of a new car today. While General Motors sells more electric vehicles in the US than any other company after Tesla, most of its lineup remains prohibitively expensive thanks to the battery pack, which accounts for 30-40 percent of the total cost.

During a tour in early September, after donning a hard hat, goggles and a protective vest, I got a close-up look at the tools GM will use to make them. Newer, cheaper batteries: large vats to mix the cell slurry, coating machines that deposit the electrode material onto the film, and machines for dozens of other steps in the process of turning minerals into cells that store and deliver energy. The factory I toured would be a low-volume factory that would allow engineers to prototype for higher-volume production than giant billion-dollar factories.

The new facility shows that General Motors is playing a long game, working to develop new cells that use less or none of the minerals controlled by China globally. Although American scientists did the original research and development on most of the new cell chemicals, Asian companies were the ones who developed, refined, and brought these chemicals into volume production.

China's 15-year effort to gain global control over the supply and processing of minerals used in today's cells, particularly cobalt, nickel and rare earths, means that future cells will have to use different minerals to escape the country's dominance. An example: GM's new LMR or lithium-manganese-rich cells use very little cobalt or nickel. Second: This spring, GM announced it was working with startup Peak Energy on sodium ion cells for energy storage.

It can take a decade or more for such new chemicals to get out of the lab and into volume production. But General Motors does not have that much time. China may sell its electric cars to the Americans by then. The company's battery engineers are confident that using new facilities inside those undisclosed buildings 15 miles north of Detroit, they can launch new cell chemistries and molds for electric vehicles and battery energy storage systems in less time than expected. The future of this company depends on its success.

[Image: An engineer inspects a press roll system at GM's new battery cell development center.

In June, General Motors battery czar Curt Kelty told me that the new Battery Cell Development Center, or BCDC, will provide "a vital bridge from the R&D lab to production" by combining virtual modeling, physical testing capabilities and small-scale production lines. The goal is to fill the missing link between creating and prototyping new cells and getting them into high-volume production.

BCDC President Mo Gallegos puts the total cost at $900 million. At peak production, BCDC may produce 4 to 5 GWh of cells. But the goal is not quantity; It is of quality

Inside the 500,000-square-foot lab, I saw much of the machinery needed to build a cell, from mixing the slurry for the coating that forms a cell electrode to the charging hardware that powers a finished cell for the first time. The small plant will be used for multiple cell chemistry and different formats, starting with prismatic cells. It aims to test industrial cell production processes, prove and optimize production techniques and recipes before implementation at scale.

While BCDC will produce actual working cells, it will do so at much lower volumes and at a slower rate than production lines in full-scale cell factories. GM battery developers then transfer the knowledge gained from this process testing to their cell manufacturing partners to bring the new designs to full volume.

These are likely to include two plants operated by Ultium Cells LLC, a cell manufacturing joint venture between General Motors and its cell partner, LG Energy Solutions. The first plant in Warren, Ohio, opened in 2022 and will make large-format nickel, manganese, cobalt and aluminum pouch cells for GM's line of electric vehicles (excluding the 2027 Bolt). The second, in Spring Hills, Tennessee, opened in early 2024 and now produces lithium-ion phosphate cells for battery energy storage.

Of the two BCDC buildings, the first has the anode and cathode on opposite sides, with the electrode assembly in the middle. A bridge transports the rolls of electrode film to a second building, where the cells are assembled, charged, and tested for 8 to 11 days. The second floor contains the required metals, films, and components, as well as a very impressive array of HVAC equipment for sterile, dust-free, and temperature-appropriate processes.

BCDC is set to produce its first test cells - LMR prismatic cells - in the fourth quarter of this year. But they won't reach volume production at Ultium's Spring Hill plant until 2028. After tests for LMR cells, sodium ions will be added. Prism cells for BESS are being developed with startup Peak Energy.

[Image: Close-up of the foil battery material before it is incorporated into the cell assembly process.

The problem the BCDC is trying to solve is the gap between creating new cellular chemical compounds on a small scale and producing them in mass quantities in the hundreds of thousands or millions. Taking them through development to commercialization means understanding how they are made. which requires testing the production processes and methods to ensure the quality of the final cells as well as their quantity.

You can think of launching a cell in four steps: developing a new chemistry, testing prototypes, figuring out how to make it, and then ramping up production.

Phase I has been concentrated in GM's Kettering R&D lab for decades. The company developed and tested new chemistry and cells for the EV1 pioneer, the 2011 Chevrolet Volt, and the LMR chemistry the company plans to launch in 2028. But it's limited to proving one cell at a time, in small "coin" or pouch formats.

GM introduced the second phase when it opened its Wallace Battery Cell Innovation Center in 2023. Wallace can make these new cells in small batches, testing whether a new chemistry continues to work and remain viable in cells similar to those used in the final products.

Now, the third phase has arrived at BCDC, right next door to Wallace. It uses real-world cell manufacturing and battery assembly machinery to make thousands of cells per day – ideally ensuring that the final scale-up in a larger factory runs smoothly. And it's built with twice the floor space originally used for future expansion, multiple chemistries, and multiple cell formats.

Historically, those second and third phases were handled by battery manufacturers. They were the missing link between new chemistry and volume production—almost always done overseas. Together, Wallace and BCDC will provide GM with approximately 1 million square feet of battery development space. (GM declined to specify the number of employees at each facility, saying it varies.)

The fourth and final phase is commercial production at scale, either in Ultium cells or other high-volume plants.

To compare the scale of each phase, Kettering R&D may produce 1 liter of slurry with which to coat the electrode film for individual cell testing. Wallace's lab takes it to 40 liters at a time. Now, BCDC will be able to produce 1,500 to 2,000 liters at a time—whereas high-volume cell factories run multiple versions of the same process at once or twice as often.

[Image: A General Motors engineer inspects equipment that gives battery developers a hands-on way to study and refine how materials work before full-scale cell production.

Before physical manufacturing begins, cell engineers use virtualization during the process. Radu Tioni, GM's director of virtual engineering for global powertrain, said the agility in developing and quickly modifying new designs comes from using "digital twins" to simulate processes before production hardware enters the process. In the automotive world, a well-known example is the use of CAD software to model structural designs and crash responses. For decades, simulations have allowed automakers to speed up their engineering and reduce the number of expensive prototype cars that must inevitably crash to validate their safety.

Similarly, from the internal behavior of the new cell chemistry to the manufacturing steps of each component—for example, how the molecules are thoroughly combined in a "double planetary high-shear mixer" to produce the slurry—the time saved through modeling can shave weeks and months off the timeline for getting a new cell from lab to market.

Recently, General Motors has relied heavily on artificial intelligence in a variety of different virtualization models. Drawing on its vast trove of 10 years of previous battery data, they said its big language models now predict the results of dozens of minor chemical changes and process changes much faster than any physical prototyping and testing can. These results are very accurate when transferred to real-world prototyping, he said.

At the chemistry level, the simulation allows GM to model "what happens if we add more silicon, what happens if you reduce the cobalt content." "We should be able to simulate something in the next few days," Kelty added. This allows the company to change production processes as needed, then remodel. If done by an overseas cell manufacturer, the process can take months.

According to them, intensive modeling of molecule mixing in slurry mixers allows GM to cut mixing time in half against the industry-recommended method, without affecting cell quality or performance. Overall, they said, the company expects to put a cell through the end-to-end process in "about half the time" — though GM declined to quantify timelines or how much time it saved. Earlier, Kelty suggested that he could allow GM to hold off on developing promising new cell chemistry for a year.

Modeling encompasses every level of vehicle development: a digital twin of the entire EV. one for the battery pack; One for the cells inside. A digital twin of the electrode that models the behavior of particles on the electrodes during cell discharge and recharge. and finally an atomic-level twin to model lithium diffusion through electrodes and electrolyte.

The methods are easily and simply transferred between chemistries. For battery cells, "the modeling is chemistry agnostic. To go from NMCA to sodium ion cells, we just change the atoms in the model. The core equations of quantum mechanics are well known," Hioni added.

[Pictured: Mo Gallegos, head of GM's new battery cell development center, with machinery that allows his engineers to prototype new cells before moving the process to large factories.

Simulating cell behavior at the atom-by-atom level requires a lot of computing power, even before simulating the entire manufacturing process to build the cell in bulk. During our tour, Gallegos noted that the Battery Development Complex draws more energy than any other part of Warren's sprawling campus.

But the company has long been doing intensive computing, starting long before the modern era of electric cars. Soon after the first emissions regulations went into effect half a century ago, General Motors began using computers to model the phases of gasoline combustion in its engines. Over time, as engines gained multiple valves, multiple cams, variable valve timing, direct injection, turbochargers, and a growing array of hardware to make the inherently wasteful combustion process more efficient, they relied on ever more powerful computer models. A specific example might be modeling the formation of soot particles in diesel combustion - and how to reduce them.

This experience and available computing power allowed the company to "unlock a lot of value" because it was "easy to use these multiphysics, highly accurate tools in many different areas that probably have never been used before," Anhayuni said.

During the tour, Hayoni said that the company's computing power is second only to the country's national laboratories in the United States. "The pace of growth in atomic modeling is rapid," Hiyuni added, indicating that the company's AI capabilities have expanded significantly in the past six months.

The combustion engine side of the house is now heavily adding AI to its toolbox, though cellular modelers seem to feel they're still ahead of the curve. According to them, the atomistic and process modeling performed in the LMR cells under development requires three times the computational capacity as the modeling of combustion processes for a gasoline engine.

Kelty is clearly energized by the possibilities of these new AI-based modeling and simulation tools. In the long term, he said, "we want to get rid of testing" — or at least greatly reduce the physical testing of prototype cells and their manufacturing methods. Less physical testing to develop and refine new cells gives GM more agility to respond to real-world test data. The ultimate goal, Kelty suggested, is for the system to develop its research and development to the point where its AI models can synthesize a new cell chemistry on its own. But they are not there, however, he said.