Nio is making a simple argument that sounds a lot better on a spreadsheet than a dependency on Nvidia: if a carmaker can design its own chip and sell the technology to others, the money starts to look less like a cost center and more like leverage. At the China Automotive Chongqing Forum on June 13, Nio vice president Ma Lin said the company had at one point spent as much as $300 million a year on Nvidia chips, a bill that would only get uglier as sales grew 40% to 50% a year.

The company’s pitch is that its Shenji chip reduces hardware cost per vehicle by about 10,000 yuan, or about $1,500, compared with a setup that used four Nvidia Orin chips in a car. That is the kind of saving that matters twice: once on the cost line, and again in gross margin, where every yuan left on the table has a nasty habit of multiplying across a fleet.

Shenji chip is moving beyond Nio-branded cars

Ma said Shenji is now being used not only in Nio vehicles but also in ONVO models this year. That matters because chip programs rarely justify themselves on vanity alone; they win when the same silicon can be spread across multiple nameplates, making the upfront R&D spend look less heroic and more rational.

Nio first announced Shenji NX9031, its own high-performance automotive chip built on a 5-nanometer process, in 2023. Since then, the company has also licensed its semiconductor technology to outside organizations, which is a neat way to turn what used to be a supplier bill into a product line.

Nio’s R&D spending is already cooling

The company said first-quarter 2026 R&D spending was 1.885 billion yuan, or $278 million, down 40.7% from a year earlier. That does not prove the chip strategy alone is responsible, but it does suggest Nio is trying to make its engineering budget do more than burn brightly and disappear.

  • Peak annual Nvidia chip purchases: $300 million
  • Estimated savings per vehicle with Shenji: about 10,000 yuan, or about $1,500
  • Shenji NX9031: Nio’s first self-developed high-performance automotive chip, announced in 2023
  • First-quarter 2026 R&D spending: 1.885 billion yuan, or $278 million

The bigger bet is control

Nio is not just trying to save money on chips. It is trying to avoid being priced by someone else, which is the real point of any serious in-house silicon program. The obvious question now is whether Shenji can keep spreading fast enough across models and partners to justify the ambition behind it, because once a company starts selling the tech outward, walking back to commodity parts gets awkward very quickly.

Source: Ixbt

Leave a comment

Your email address will not be published. Required fields are marked *