BYD Didn't Build Batteries for Its Cars. It Built Cars for Its Batteries.

Pull a BYD Seal apart and weigh what's left. When UBS engineers did exactly that in 2023, about three-quarters of the car - cells, packs, electronics, structure - had been made by BYD itself, against roughly half for a made-in-China Tesla Model 3.¹⁵ The cars look like competitors on a showroom floor. Under the skin they are two different organisms. One is an assembler buying its most expensive part from a rotating cast of suppliers. The other grew the part first, then grew a car around it.

The story everyone tells is that BYD is a carmaker that got clever and integrated backwards into batteries. Almost every word of the framing is the wrong way round. BYD did not start with cars and reach back for cells. It started with cells and reached forward for cars - and that single directional fact is the whole cost story.

It was a battery company before it was anything else

BYD was formally registered in February 1995 as a battery manufacturer - rechargeable cells, the kind that went into the phones of the late 1990s.³⁴ It did not build a single car for the better part of a decade. The carmaker only exists because, in January 2003, the battery company bought an automaker - the struggling Xi'an Qinchuan - and pointed its existing cell business at a new product.³ That sequence matters more than any clever engineering that came later. A company that learns to make batteries first and cars second carries the cheapest, hardest-to-copy part of an electric vehicle in its DNA, not on a purchase order.

By the time the cars dominated the business, the inversion was complete. Automotive crossed half of BYD's revenue back in 2009 and passed 80% by 2023⁸ - but the engine underneath was still a battery maker. Its cell subsidiary, FinDreams, is now the second-largest EV battery producer on earth behind CATL.⁵ BYD is, in effect, a giant battery company that happens to wrap most of its output in sheet metal and sell it as cars.

Why the direction decides the cost

Here is the part that the 'BYD is just efficient' narrative misses. In an electric car, the battery is not one component among many - it is the single most expensive thing in the vehicle. Whoever owns the cell owns the cost curve. Tesla, for all its software brilliance, still buys most of its cells from a rotating bench of suppliers: Panasonic in Nevada, LG in China and Europe, CATL in Shanghai - while ramping its own 4680 line in Texas, which passed 100 million cells by September 2024.⁷ That is a hybrid, and a hybrid still pays a supplier's margin on the part that matters most. BYD pays itself. It mines the input, makes the cathode, casts the cell, builds the pack, writes the management system, and bolts the whole stack into a body it also tooled.⁷ Every handoff that would normally carry a markup happens inside one balance sheet.

That ~15% per-vehicle gap is why the comparison to a software lead is so misleading.¹ A software advantage can be coded around; a rival can hire, copy, or partner its way to feature parity in a model cycle or two. A vertical-integration cost advantage is a physical thing - plants, processes, materials contracts, and a quarter-century of battery learning that you cannot acquire on a weekend. UBS put BYD's edge over legacy European makers at around 25%.¹¹⁰ The moat isn't a clever app. It's a supply chain that starts in a mine and ends on a dealer lot, with no one taking a cut in between.

The honest objections: subsidies, and a margin that isn't magic

Two fair counters deserve to be on the table, not buried. First, the state. Pro-BYD integration stories love to credit pure operational genius and skip the fact that BYD took hundreds of millions to billions of euros in Chinese state subsidies across the early 2020s - rising sharply through 2022, per its own annual reports. Some of the cost edge was funded, not earned. Any honest read has to concede that the moat was dug partly with public money. Second, the margin. Vertical integration does not magically make BYD wildly more profitable than Tesla - the UBS teardown pegged the Seal at a 16% gross and just a 5% EBIT margin, roughly mainstream-car economics, not luxury-tech economics.¹ So the claim is narrower than the cheerleaders want: BYD's advantage is in cost per car, which it can choose to spend on price, not in a fatter headline profit.

But notice what survives both objections. Subsidies helped build the plants; they don't run them today, and a subsidy cannot teach a rival twenty-five years of cell chemistry. And a modest EBIT margin on a structurally cheaper car is precisely the point - it is the ammunition. BYD can hold a thin margin and still undercut everyone, because its floor is lower than theirs. The state lit the fire; vertical integration is what keeps it burning cheaply.

The numbers the inversion built are no longer subtle. In 2024 BYD's revenue topped RMB 777 billion and crossed $100 billion for the first time - past Tesla's $97.7 billion - on 4.27 million vehicles, with R&D spending of RMB 54.2 billion that ran higher than its own net profit.²⁶ A battery company that learned to make cars is now out-earning the carmaker the world treats as the future. The tempting read is that BYD is a Tesla that got cheaper. The truer one is that it was never trying to be Tesla at all. It built cars the way a foundry builds anything - from the raw material up, owning the part everyone else has to go buy. You can copy a feature. You cannot, on short notice, become a battery company that has been one since 1995.