GM’s big new battery tech push

The cost logic is a repeat of a move the industry already made once. In the nickel-manganese-cobalt family, early cells used the three metals in equal proportions, which the transcript renders as “1, 1, 1,” and because cobalt was the most expensive, the whole industry spent years driving cobalt out, arriving at high-nickel cells that have climbed from 60% to 70% to 80% and now as much as 90% nickel, with a few percent each of cobalt and manganese. Kelty’s description of LMR is the same trick applied one metal along: take out most of the nickel, which is now the expensive input, and fill the space with manganese, which runs around $2 a kilogram. The penalty is energy density, and he states it plainly rather than burying it. His illustration is the Chevrolet Silverado electric truck, which he puts at over 490 miles of range on high nickel. Keep the same pack and fit lithium iron phosphate and you get about 350 miles; fit the new chemistry and you get about 400, at lithium iron phosphate pricing per watt-hour. That middle slot is the entire product thesis.

What makes the claim more than a pitch deck, in his telling, is that nothing around the cell has to change. GM says LMR runs on the same Ultium factories with the same electrode manufacturing and the same packaging, with only the formation step differing and the equipment unchanged, a claim he qualifies with “unless you change the form factor.” Upstream matters more: the same suppliers that make high-nickel cathode can make this cathode on the same equipment, there is surplus cathode capacity in the market, and a couple of process steps actually drop out, which makes the material cheaper to produce. The anode is still graphite and there is still about 1% or less cobalt. That is the claim worth understanding, because building a new factory and a new supply base is where most new chemistries die. The idea is not GM’s; Kelty traces LMR to Jeff Dahn’s lab at Dalhousie University about 20 years ago and then to Argonne, after which it sat parked because nobody solved the technical problems. GM names two it says it solved: cycle life, which initially was not good enough, and formation time, the finishing step before cells ship, where extra hours translate directly into cost. The evidence offered is internal and partner testing. GM built full-size cells at an in-house large-cell center it opened about two years earlier, tested them under the conditions a vehicle would see, then ran the chemistry on partner LG Energy Solution’s pilot line before committing.

Kelty’s own framing is that every battery announcement hides a missing metric, and he claims LMR has no drawbacks, but he scopes that claim immediately when Kann pushes on stationary storage. A vehicle needs roughly 1,000 to 1,500 cycles over its life. Storage cycled daily needs about 3,600 cycles over ten years and twice that over twenty, and there lithium iron phosphate remains the preferred chemistry. He allows LMR into storage only where cycling is occasional, on the order of 30 to 50 times a year, where the smaller footprint pays. So the honest version of the claim is chemistry by segment rather than a single winner: high-nickel for premium long-range vehicles, LMR for the price-sensitive middle and specifically for trucks and full-size sport utility vehicles where the pack has volume to spare, lithium iron phosphate at the low end, all three coexisting for at least five years. When Kann asks the sharpest question, whether LMR shares lithium iron phosphate’s eventual price floor and whether it can compete with the very cheap Chinese electric vehicles built on that chemistry, Kelty answers with the segmentation argument and does not address the price floor.

The supply-chain half of the conversation sets the constraints all of this sits inside. Packs and modules have to be built near the vehicle because shipping them is prohibitive, while cells travel better; GM says all cells for its 12 electric vehicles come from its North American plants. Upstream is Asia-weighted: graphite for the anode almost entirely from China, cathode currently from Korea, precursors and metal sulfates in Asia, nickel mainly from Indonesia, cobalt from the Democratic Republic of Congo. Absent tariffs and subsidies, Kelty argues, rational actors process at the mine, because shipping rock makes no sense and nickel sulfate is mostly liquid, so the powdered precursor is what travels and cathode production centralizes somewhere convenient. He says making the cathode in North America already makes economic sense, and that GM is working with partners to do it. Making the precursor here is the harder case: he expects it would most likely cost more, because environmental rules are tougher and the ore is not necessarily as concentrated. His caution about the policy fix is the interesting part: tariffs plus incentives can pull the supply chain onshore, but they raise costs and therefore vehicle prices, which suppresses demand and defeats the purpose, and the production credits expire around 2031 or 2032, so whatever gets built has to be economic without them. He says GM is still working out the optimal strategy.