It wasn’t supposed to be like this.
By early 2024, General Motors was to have had half a dozen electric vehicles selling in volume, all using the next-generation Ultium architecture it unveiled at “EV Day” in March 2020. That list includes the Chevrolet Silverado EV, Blazer EV, and Equinox EV; the Cadillac Lyriq; and the GMC Hummer EV in two versions. This expansive rollout of new EVs was to put GM on a course to, as The New York Times put it in 2022, “challenge Tesla for the lead in EV sales before the end of the decade.”
Yet, by last December, they were all selling at minimal rates or outright unavailable. One model has been under a stop-sale order for months. Others are severely behind schedule. While many traditional and startup automakers are struggling with the EV transition, GM’s electric future seems especially stalled right now.
That doesn’t mean it’s DOA. As of today, the Hummers, the Lyriq, and a Work Truck version of the Silverado EV are on sale. About 500 Blazer EVs were sold before deliveries were paused in December. The Equinox EV is still set to arrive this year. GM continues to reveal further Ultium models from Cadillac and GMC. Another Ultium EV, the 2024 Honda Prologue, will go on sale next month. GM builds that vehicle for Honda in its Ramos Arizpe assembly plant alongside the Equinox EV; it will do the same with the Acura ZDX that parallels the Blazer EV and Cadillac Lyriq.
All of that means that this year may be a make-or-break one for GM’s electric game. But understanding how it got to such a low point helps illustrate the many challenges ahead.
GM's Absent Ultium Cars
This article was originally entitled, “GM’s EVs MIA After 18 Months: WTF?” Cooler editorial heads prevailed. But now we can add detail to the strange saga of what’s been going on with the plan by Detroit’s largest carmaker to sell high volumes of battery-electric vehicles.
We spoke with sources deep in the EV battery world—who have worked with multiple automakers—to ask what they knew about the Ultium problems. In all, GM has had to overcome significant hurdles not only with the assembly of Ultium battery modules, but also, separately, in software for both vehicle charging and a new infotainment system that eliminates Android Auto and Apple CarPlay.
The numbers are stark. Two years ago, in February 2022, GM said it expected to sell 400,000 EVs by the end of 2023. Eight months later, it rolled that date back to mid-2024. (In October, it abandoned the 400,000 number altogether.) But its actual 2023 sales of Ultium EVs were a mere 13,838. Of those, 9,155, or two-thirds were one single model: the Cadillac Lyriq. Sure, GM sold 76,000 electric cars altogether, but a whopping 82% of them were the aging Chevrolet Bolt EV and EUV—which it took out of production late last year.
The low Ultium sales don’t stem from a lack of buyers. Despite recent grim headlines suggesting otherwise, U.S. electric vehicle sales continue to grow. Last year’s were the highest ever, in volume and market percentage. Instead, GM simply hasn’t been able to build enough Ultium EVs to get the needed stocks to dealerships. Starting last summer, when the Blazer EV began to ship to dealers, GM executives have repeatedly had to explain and make excuses for the delays.
Two points suggest those delays may be ending. First, John Hwang, the Prologue product lead for Honda, said he is confident Honda will receive 40,000 Prologues from GM during 2024, and 70,000 a year starting next year. Given the dismal Ultium sales in 2023, those totals alone are a big jump. Second, until now, GM has been silent about the specifics of the problem beyond high-level statements by CEO Mary Barra. This month, InsideEVs was able to get more details from Mike Anderson, GM vice president for global electrification and battery systems.
Perhaps unexpectedly, Anderson confirmed the vehicle delays don’t stem from Ultium Cells LLC, the GM-LG joint venture company that fabricates Ultium cells. Getting a new cell plant up to usable yields is a grueling process that takes years to reach profitability. Panasonic, one of the world’s most competent battery makers, produced the first cell at its Tesla Gigafactory plant in Nevada in January 2017, but even it said the factory didn’t turn profitable until Q1-2021.
So you might expect cell supply from the Ultium Cells LLC plant in Warren, Ohio, would be the problem—but that doesn’t appear to be the case. Last year, Anderson said, the Ohio cell plant “exceeded some of their operational efficiency targets” and “did a great job. Man, they killed it.”
'A Box With Sides'
Instead, GM has said the delays stem from problems with the automated machinery that assembles the large 9-kilowatt-hour modules that go into the battery pack of an Ultium vehicle. Anderson described the assembly process this way:
Think of it as a box, it’s got six sides. The bottom of the box is a cooling plate. There are 24 cells in a box, these long pouch cells, and they’re attached to the cooling plate with TIM (thermal interface material)—think of it as epoxy and a way to transfer heat away from the cells. Then there are five other sides you have to put in. The top has two ICB (interface control board) frames that are hinged to it. The cell tabs actually go through those frames and they get bent and welded on. The sides are put on with a structural adhesive material.
The automation equipment does all of that automatically. You push the parts of the box and the cells up to the end of the line, robots grab the cells, they go on gravity tables, and everything is auto-cut, auto-bent. We built mini-stacks of cells (eight in series, three in parallel), then we build a mega-stack, and it all comes together.
The box is put together in what we call a press. All these parts are coming together and they have to press together just right, the glue has to be just right, everything fits in the grooves, everything snaps into place. It’s been a learning exercise for all of us. Everyone does it their own way, but no one does it this way.
The problem, Anderson said, was that this process was very, very difficult to get up to speed. The company started each of its module assembly lines at five jobs per hour, with the goal of ramping up to 30 per hour in early 2023.
And that just didn’t happen, despite months of what he said were 24-hour days to revise specs, modify drawings, adjust tolerances, ensure more precise alignments, better control the flow and timing of the glue, and all the other steps toward “getting that to run at those rates, with quality, making sure they’re perfect at the end.”
The Mystery Integrators
The Ultium architecture is the first time GM has developed a battery system on its own.
While they were assembled in a GM plant, the batteries for both the earlier Chevy Volt plug-in hybrids and the Chevrolet Bolt EV were largely developed by cell provider LG. Chevrolet engineers worked alongside them, and LG designed battery packs to the packaging and performance specs GM provided—but Chevy didn’t do assembly or the crucial systems integration of those batteries.
The Ultium vehicles, on the other hand, are largely GM’s work. While it created Ultium Cells LLC as a joint venture company with LG Energy Systems to fabricate the cells, GM designed the modules that hold the cells, the battery packs that hold the modules, and the wireless battery-management system. Suppliers manufacture the physical components—Magna for battery cases and Lear for the battery-disconnect unit, for instance—but GM does the assembly and integration.
One source within the battery industry suggested to InsideEVs that GM ordered its automated module assembly system from a traditional Tier 1 supplier with little experience in high-volume electronics assembly, rather than a firm that had already built module lines for other EV makers.
GM declined to identify its supplier, or even to characterize the type of company it is. Under repeated questioning, Chevrolet spokesperson James Cain would say only that it was “an industrial equipment supplier that works across multiple industries, not just automotive.” Anderson added little more, saying only that it was a company with a “background in automation” with experience buying subsystems like conveyors, robots, and welders, and integrating them into a complete system for bringing together complex assemblies.
“But nobody has experience doing this job; I’ll put it that way, right?” Anderson laughed. He declined to name names or “call out any particular suppliers or anything like that.” He also noted GM has more than one automation supplier or integrator. But, he stressed, “No one’s really done it at this rate and this scale, including ourselves.”
Repeated Excuses; Barra 'Disappointed'
After what was expected to be a gradual ramp in 2022, last year was intended to be the big year for Ultium volume. That didn’t happen. And despite many headlines highlighting its EV delays, GM has been exceptionally cagey about what went wrong. Discussing Q2 results in July, Barra said:
We have experienced unexpected delays in the ramp-up because our automation equipment supplier has been struggling with delivery issues that are constraining module assembly capacity. We are working on multiple fronts to put this behind us … we have deployed teams from GM manufacturing engineering to work on-site with our automation supplier to improve delivery times … we have added manual module assembly lines.
To compensate, she said, the company would install more module capacity at all of its North American EV plants. Those include Factory Zero (formerly called Detroit-Hamtramck); Spring Hill, Tennessee; Ramos Arizpe; and the CAMI plant in Canada.
In October, very little further information emerged from the third-quarter results call:
Our battery module constraint is getting better, which helped us more than double Ultium Platform production in the third quarter compared to the second quarter. We are now in the process of installing and testing our high-capacity module assembly lines, which will continue into the first part of next year. We are currently challenged with getting some of the critical equipment components. We have a dedicated team working with our suppliers to resolve all issues to get these lines running at rate.
During a November business update conference, Barra said she was “disappointed with our Ultium-based EV production in 2023.” She said the company expected “significantly higher Ultium EV production” in 2024. When might that happen? “By mid-year [2024], we expect modules will no longer be a constraint, and we will focus on building to customer demand rather than setting new production targets,” she said on the October call.
GM representatives suggest the worst is past. Anderson said Cadillac sold more than 2,000 Cadillac Lyriqs in December and another 1,900 in January—with the January number representing 19% of the brand’s sales that month. For Q4, Hummer sales ran at about 700 a month and Silverado EV sales at 450 a month.
Importantly, GM has put its Ultium vehicles into production in China with little difficulty. Those cars use different battery cells than the North American versions, either one of two nickel-manganese-cobalt cells (“thin” and “thick”) or lithium-iron-phosphate “blade cells” from Chinese powerhouse BYD. Anderson confirmed that a different company proved the module-assembly equipment for those batteries.
Since China has a majority (57%) of the world’s EV sales, with EVs making up more than one-third of last year’s new-vehicle sales, it likely now has more and more capable integration vendors and more experienced engineers and managers to oversee such hardware. As Anderson said, “This is a major capacity burst for the whole [North American] industry, so everyone who makes this kind of equipment is saturated right now.”
Asked what he might change if he got a do-over, Anderson said, “I think where we got challenged, besides it being new tech, was when we decided to really accelerate the ramp. The combination of the two challenged even our experience, right?”
What We Might Assume
GM’s problems getting Ultium launched may have been compounded by the departure of many of the company’s most experienced battery engineers through retirements, buyouts, and even poaching by other firms. Anderson himself has been at GM 33 years, since his 1990 graduation from Purdue.
And while the most experienced battery engineers in the U.S. largely come from Silicon Valley—largely meaning “Tesla”—GM has traditionally been resistant to importing outside expertise.
The October resignation of Gil Golan, just one month after the executive from Israeli battery startup Algolion was appointed GM’s CTO (GM had bought the company in July), added to the turmoil. In a profile in The Electric, he had expressed his desire to bring in senior battery experts from outside the company. Management may have believed that wasn’t needed, that GM could darn well do the job itself with its existing staff. Contrast that to Ford CEO Jim Farley, who hired ex-Tesla executive Doug Field from his position overseeing the secretive Apple Car project in September 2021 when he felt in-house EV engineers needed more expertise and fresh thinking.
Just recently, GM may be reversing course there. Earlier this month it hired Kurt Kelty, a longtime former Tesla battery expert, to ramp up its long-term production apparatus and bring its second-generation Ultium system into being without the delays the first has experienced. For a company famous for its insular culture, it’s a huge move—and maybe an admission that it needs help playing a new ball game.
That's Not All, Folks
Its inability to get the Ultium vehicles into production is far from GM’s only problem. It was also outsold in EVs in 2023 by a smaller rival, and its sales in China have plummeted. Journalists, analysts, and customers alike question its decision to eliminate Android Auto and Apple CarPlay from its future vehicles in favor of a GM-developed alternative pairing method.
First, GM was outsold last year in EVs not just by Tesla, but also by a somewhat unlikely rival. That would be Hyundai-Kia, which remains the company to watch in EVs alongside Tesla and a few aggressive Chinese makers like BYD. A new Bloomberg NEF analysis says Hyundai-Kia accounted for more than 8% of the passenger EVs sold in the U.S. during 2023. Only Tesla has a (much) higher percentage. The Korean company’s less-expensive entries (Hyundai Kona EV, Kia Niro EV) neatly fill the niche GM abandoned when it killed off the Bolt last year.
Second, and worse for its long-term prospects, GM’s sales in China have plummeted to just half what they were six years ago. Buick and Cadillac continue to sell more units in China than in North America. But sales of volume brands built with joint-venture partner SAIC have fallen substantially as the market shifts toward homegrown Chinese brands—Geely, BYD, Great Wall—whose EV models are now fully competitive with foreign entries, and which include more advanced software systems than Western makers can muster. GM’s response has been to move its products upmarket in China to be, in Barra’s words, “more premium and high-end,” including the Ultium-based Cadillac Celestiq starting at $300,000. The results of that pivot are yet to be seen.
Finally, GM stunned the industry when it said it would eliminate Android Auto and Apple CarPlay, standard equipment today on every new vehicle sold. It claims its in-house alternative will provide all the same functionality, but early reviews and tests suggest it may not yet be fully baked. Similar problems also produced a stop-sale on its new gas-powered 2024 Chevrolet Colorado and GMC Canyon midsize pickup trucks. Barra said last week at a global auto-tech conference in New York that among her regrets from last year was not bringing in GM's new software team sooner.
Predictions Are Risky
While GM continues to be a very profitable company, its stock price remains at the same level (not accounting for inflation) as when Barra took the reins 10 years ago this month. Any experienced reporter also knows it’s always dangerous to count out a major automaker.
As of this month, Anderson said GM was continuing to build its various Ultium vehicles—even if the problematic Blazer EVs had to be stored temporarily until their software could be updated. “We’re not losing the opportunity to make product,” he said. “We’re not losing the build spot. If you’re going to make [those] kinds of numbers, you can’t take days off, you can’t get time back, so we’re not losing the time. You can’t take time off and then say, ‘I’m gonna go twice as fast’.” As with the midsize pickups, once the software problems are identified and fixed, Ultium vehicles that have been stored will go through a cleaning and validation process before they are sent to dealerships.
A GM representative told InsideEVs it hopes to produce 200,000 to 300,000 Ultium-based vehicles for GM brands this year. Add to that 40,000 Honda Prologues—about as many Mustang Mach-Es were sold last year—and you’re looking at a quarter of a million or more Ultium vehicles this year—which would boost EV market share at least 2% over last year’s total in the U.S.
Meanwhile, GM also said it would resume offering plug-in hybrid models in North America, which it hasn’t done since it killed the Chevrolet Volt after a truncated 2019 model year. It still offers PHEV powertrains in other markets like China, though it hasn’t specified which U.S. models may become PHEVs as yet. It remains to be seen whether these will serve only as interim compliance cars to keep GM on the right side of emission rules, or whether the company will actually make substantial efforts to explain how PHEVs work to a largely uncomprehending mass market.
GM’s first-quarter sales, due out in early April, will prove or disprove the company’s confidence. And we’ll continue to watch closely to see when and which actual Ultium vehicles arrive in volume at dealerships—and how effectively GM’s “valued partners,” the independently operated third-party franchised dealers, choose to sell them.
For GM’s sake, those Ultium vehicles had come soon—and they’d better work properly right out of the box.
John Voelcker covers advanced auto technologies and energy policy as a reporter and analyst, specializing in electric vehicles and the energy ecosystem around them. He edited Green Car Reports for nine years, and his work has appeared in Car and Driver, The Drive, Forbes Wheels, Wired, Popular Science and NPR's "All Things Considered."
