Meituan is best known in China for getting dinner to your door fast.

Now its investment arm is leading a round for a computer chip that hasn’t been manufactured yet.

Shanghai Lingrui Zhixin Computing Technology, known as Lingrui Zhixin , announced on November 13, 2026 that it has closed a Pre-A+ round worth several hundred million yuan. Meituan Strategic Investment led it. The startup didn’t share an exact amount or a valuation.

A Pre-A round is early money, raised before a startup’s first big institutional Series A. The plus sign means this is an extra round at that same stage.

The rest of the investor list is long, and it’s full of companies with their own reasons to care about computing power. There’s a fund affiliated with JD.com, SenseTime’s Guoxiang Capital, Shangqi Capital (the venture arm of carmaker SAIC) and Amber Capital, the industrial fund of network security company Sangfor. Ganbei Fund, Junxi Capital and the Lihe Suanxin Fund joined too, along with many existing shareholders.

Liu Yang’s startup, not even two years old

Lingrui Zhixin was founded in early 2025. It’s based in Zhangjiang, the chip-heavy corner of Shanghai’s Pudong district, and its founder, chairman and general manager is Liu Yang.

With this round, the company says it has raised close to 1 billion yuan in total.

In under two years.

In July, the startup announced a round of several hundred million yuan from backers including Zhangjiang Hi-Tech, CAS Star and the GPU maker Biren Technology. Four months later, Meituan walked in.

Money for AI hardware has been moving quickly outside China as well. In the UK, chip startup Fractile raised £577 million in its Series C.

The P100 bet

Lingrui Zhixin designs CPUs, the general-purpose processors that handle a computer’s main work. It builds them on RISC-V.

RISC-V is an open instruction set, the basic list of commands a processor understands. Anyone can use it without paying license fees, unlike x86 or Arm, and Chinese chip companies have piled into it.

This year the company released the P100, a server-class CPU core. A core is a single processing unit inside a chip, and a server chip can pack many of them.

The feature Lingrui Zhixin talks about most is SMT4. Simultaneous multithreading lets one core work on several streams of instructions at the same time, and the P100 handles up to four, switching between one and four as the load changes. The company says no other RISC-V core in China does this. Why does it matter so much to them? Because Lingrui Zhixin is building for AI agents, the software assistants that break a job into steps, call tools, wait for answers and keep going. Agents throw off a lot of small tasks at once, and they need those tasks finished quickly and reliably. The company chose to design its CPU around that kind of crowded, many-at-once workload, and it calls the result “agent-native.”

It lists data centers, edge computing (processing done close to where data is collected), robots and self-driving cars as places the P100 could end up.

Wait, the delivery app wants a say in CPUs?

That was my first reaction too. Then again, AI computing keeps pulling in unlikely names. Torus, a flywheel company born on a tree farm, raised $500 million to start selling it.

So I looked at who else signed on here.

Meituan runs one of China’s busiest apps. JD runs a huge online store with its own logistics network. SenseTime builds AI models. SAIC makes cars. Sangfor sells security and networking gear. Each of them either burns through serious computing power or ships products with chips inside.

Seen that way, the investor list reads a lot like a list of possible customers. Liu Yang more or less said so. He described working with the industrial shareholders on where the chips get used and how they get rolled out.

Meituan’s strategic arm had already backed Reexen Technology, a compute-in-memory chip company. Lingrui Zhixin is its newest chip bet.

None of these investors

Laura is a business historian and researcher who writes about how companies began and grew. At Companies History, she works primarily from original sources such as annual reports, regulatory filings, patents, and archived news coverage. Her articles often correct popular myths about famous founding stories.