ECARX and Arm sign MoU to explore high-performance AI computing for vehicles and physical AI
ECARX and Arm have signed a non-binding memorandum of understanding, announced on November 2, 2026, to jointly explore next-generation high-performance AI computing platforms for mobility and cross-industry physical-AI applications — building on SiEngine chip work that has produced millions of 7nm SE1000 units since March 2023, with no new chip announced.
- Who and when: ECARX and Arm announced the agreement on November 2, 2026, in a release issued from London.
- What: A non-binding memorandum of understanding to jointly explore and develop next-generation high-performance AI computing platforms.
- Joint work: System-level architecture, performance optimisation, software enablement, functional safety, toolchain and cybersecurity validation, and ecosystem build-out.
- Where it would run: Tailored configurations for China and global markets, for mobility and broader cross-industry physical-AI applications.
- Precedent: SiEngine's 7nm SE1000 automotive-grade chips have been produced and deployed in the millions since mass production began in March 2023; the 5nm SE2000 platform is under development.
- Scale and unknowns: ECARX says its technology is deployed in more than 12 million vehicles, with 18 automaker partners and 28 vehicle brands. No chip, specifications, financial terms or timetable were disclosed.

ECARX and Arm announced on November 2, 2026 that they have signed a non-binding memorandum of understanding (MoU) to jointly explore and develop next-generation high-performance AI computing platforms for mobility and broader cross-industry applications. The announcement, issued from London, frames the work as an extension of a long-standing collaboration between the two companies — not as a product launch.
Under the MoU, the companies will pursue joint co-engineering spanning system-level architecture definition, performance optimisation, software enablement, functional safety, toolchain and cybersecurity validation, and ecosystem build-out. They will also explore tailored configurations optimised for deployment across China and global markets, covering use cases in mobility alongside broader physical-AI applications. Arm will support ECARX in bringing highly optimised compute solutions to market that are purpose-built for demanding physical-AI workloads, AI Affairs reported.
What “physical AI” means here
Physical AI is the industry’s shorthand for AI that acts in the physical world — in vehicles and other autonomous systems, rather than in a chat window. Arm’s Drew Henry, executive vice president for physical AI, called it “one of the most complex systems engineering challenges our industry has taken on, requiring compute, software and safety to work together from the ground up.” ECARX chief executive Ziyu Shen said the companies’ teams “will now advance joint technical roadmap definition and co-engineering to translate this shared vision into practical solutions for global markets.”
The starting point is the AI-defined vehicle: cars whose capabilities are increasingly set by software and central computing rather than by fixed hardware. The MoU extends that brief beyond cars, to other industries where machines need to perceive, decide and act — though neither company named a specific non-automotive product or customer.

A track record in car chips, not a blank sheet
The agreement builds on work the companies have already shipped together through SiEngine, an automotive semiconductor company. ECARX said SiEngine has delivered advanced 7nm SE1000 platform automotive-grade chips to global automakers, with millions of units produced and deployed since mass production began in March 2023. Development of the next-generation 5nm SE2000 system-on-chip platform is also under way.
ECARX brings an existing automotive footprint to the joint work: the London-headquartered company, listed on Nasdaq as ECX, says its technology is deployed in more than 12 million vehicles worldwide and that it works with 18 global automakers and 28 vehicle brands. Founded in 2017 and listed in 2022, it operates from 13 major locations with a team of more than 1,400 employees. That installed base is the closest operating comparison for the proposed computing work — the MoU itself announces exploration and product definition, not a shipped platform.

What the MoU does not do
Because the memorandum is non-binding, it commits neither company to a product, a specification or a delivery date. No new chip was announced, no performance or power-efficiency figures were published, and no financial terms or timetable were disclosed. The test of this agreement will be whether the joint technical roadmap and co-engineering Shen described turn into defined products — not the signing itself, which is a framework for exploring them.
Frequently asked questions
What did ECARX and Arm actually agree to?
They signed a non-binding memorandum of understanding, announced on November 2, 2026, to jointly explore and develop next-generation high-performance AI computing platforms for mobility and broader cross-industry physical-AI applications, through joint co-engineering and product definition. No product was announced.
What chip work does the agreement build on?
Their existing collaboration runs through SiEngine, whose 7nm SE1000 automotive-grade system-on-chip platform has been produced and deployed in the millions since mass production began in March 2023. Development of the next-generation 5nm SE2000 platform is under way.
Is this a product launch or a chip announcement?
No. The memorandum is a non-binding framework for exploration, joint research and development, and product definition. ECARX and Arm disclosed no new chip, specifications, financial terms or timetable.