The Ban That Bites Its Lobbyists: Severing Chinese Components Lifts US EV Costs and Squeezes Ford/GM/Tesla Margins

On September 3, the Alliance for Automotive Innovation — whose membership includes General Motors, Ford, Toyota and Stellantis — wrote to congressional leaders demanding a permanent statutory ban on Chinese vehicles, software and hardware before the 119th Congress adjourns. The lobbying is the setup, not the story. The story is that the same members are the industry's largest carriers of the very componentry the ban would sever. GM, the bill's most visible corporate backer, reports roughly 20% Chinese content in three of its best-selling Chevrolet models. Toyota's plug-in Prius runs about 15%. The weapon the industry is handing Congress is aimed at its own supply chain.

The test counts ownership, not technology

Yet the ban's teeth are narrower than its language, and that is the decisive detail. The Connected Vehicle Security Act of 2026, which cleared the Senate Commerce Committee on July 22, and the January 2025 Commerce rule it builds on are keyed to ownership and control, not to where a technology was invented. The bill bars from the US market any company more than 15% owned by Chinese entities. The Bureau of Industry and Security rule prohibits transactions in Vehicle Connectivity System hardware and software that is "owned by, controlled by, or subject to the jurisdiction of" China or Russia. Neither test asks which country's labs produced the chemistry or the code.

That yardstick draws a line straight through Ford's battery plant in Marshall, Michigan. Ford owns and operates BlueOval Battery Park; CATL licenses the lithium-iron-phosphate chemistry, the manufacturing process and the staff training. There is no Chinese equity to flag and no board seat to vote. When Transportation Secretary Sean Duffy wrote CEO Jim Farley in September to demand Ford reconsider, calling the arrangement grounds for "profound concern", Ford answered with structure rather than contrition: the plant is owned, controlled and staffed by Ford, and CATL is a licensor. Under a test that measures ownership, the license is invisible. This is the mechanical reason Ford can tell lawmakers it will cooperate with a Chinese-car ban and keep licensing the cheapest battery cells on earth inside an American factory wall.

The price the severance is meant to exact

The value of that license is the cost gap that a severance would expose. CATL has been selling LFP storage cells at about $63 per kilowatt-hour online and has claimed a large-format cell under $56. The global average LFP pack runs $81/kWh; the average across all chemistries is $108; nickel-manganese-cobalt runs $128. On a 70 kWh pack, a maker cut off from Chinese cells and forced onto the global LFP average pays on the order of $1,200 to $1,500 more per vehicle before it re-engineers chemistry into a substitute's form factor and thermal behavior — an estimate, but the direction is not in dispute. No US-based producer is close to closing that gap this decade. The Chinese cost advantage is exported scale and subsidized capacity, not a production-cost level an American plant converges on overnight.

Ford's structure circumvents the battery node entirely; that is the point. But the ban is not water at the nodes where the exposure is ownership rather than license, and there the margin squeeze is real. The Commerce rule's "component contamination" logic is draconian: if a single software subcomponent of an automated-driving suite comes from a Chinese entity, the entire suite is treated as foreign-adversary supply, so one flagged chip forces a redesign of the whole stack. The supplier base beneath US vehicle badges is more Chinese than the branding suggests — more than 60 US auto suppliers are owned by companies in China, and Chinese shareholders hold stakes in roughly 10,000 American suppliers. Chinese parts routinely cost about half of their US equivalents. The 2027 software and 2030 hardware cutoffs, plus GM's own 2027 supplier deadline, force these swaps on a calendar, and nothing licenses the exposure away, because the exposure is ownership.

The exposure screen, node by node

The falsification test tells you how real the "bites its lobbyists" reading is. The severance thesis is not water-tight: at the battery node the ownership test leaves a licensed channel open, so the largest single dollar exposure — Ford's cells — never gets severed at all. But it equally cannot be neutralized by substitute supply: the components where the ban does land have no cheap non-Chinese alternative and no licensing loophole. The result is a per-node screen rather than a per-ticker verdict. Ford keeps its low-cost LFP through a license and absorbs the hardware shock on its connectivity stack; GM, having already demanded suppliers strip Chinese parts, eats the re-engineering cost across a 20%-Chinese-content vehicle line; Tesla, which has told suppliers to exclude China-sourced parts, pays for what the others are losing access to. All three pushed a ban whose hardware cutoffs land on the shared bills of materials beneath their own fleets — and the one seam the ownership test cannot close is the license Ford kept.