PsiBot: $1.48B Valuation as Chery Bets on China's 22nd Embodied AI Unicorn

PsiBot raised close to $100 million at a $1.48 billion valuation led by Chery Automobile. That's the short answer. The longer answer is more interesting.

Bloomberg reported the deal on July 23, 2026, and it's the kind of round that would have been the biggest robotics story of the month in any other year — except this is China's embodied AI sector in 2026, where a $1.48 billion valuation makes you the 22nd unicorn minted this year, not the first. What makes PsiBot worth a closer look isn't the number, it's who's writing the check and why: an automaker, not a traditional VC fund, betting that the valuable part of the humanoid robot isn't the body — it's the brain.

PsiBot $1.48 Billion Valuation: Round Terms and Lead Investors

PsiBot, also known by its Chinese name Lingchu Intelligence, is finalizing a round of close to $100 million at a $1.48 billion valuation. Chery Automobile — China's largest passenger-vehicle exporter — is leading the round, joined by Lens Technology, the precision glass and components supplier that counts Apple and Tesla among its customers. The Beijing-based startup, founded in 2024, has now raised roughly $300 million in total, following a $280 million angel-and-Pre-A round disclosed in March 2026 that pulled in state-backed capital including Guokai Finance, Guozhong Capital, a CCTV-affiliated industry fund, and Shanghai's Xuhui Capital.

Why a Car Company Is Leading a Robotics AI Round

Chery leading this round is the actual story. Automakers don't typically write growth-equity checks into AI software companies — they buy from suppliers, or they build in-house. Chery's decision to lead instead of license reflects a specific read on where embodied AI value accrues: manufacturing automation today, in-vehicle and industrial robotics tomorrow, and a first-mover claim on PsiBot's models before a competitor locks up exclusivity. It's the same logic that pulled Lens Technology — a components supplier with no obvious reason to own equity in a robotics-brain startup — into the same round. Chinese industrial capital is increasingly investing directly in the software layer underneath robotics rather than waiting to procure finished hardware off a shelf.

That pattern matters for how you should read the round size. A $100 million check from strategic industrial investors isn't the same signal as $100 million from a multi-stage VC fund chasing a financial return — it's closer to a supply-chain hedge, priced as equity. Whether that discipline holds up if PsiBot's models underperform is the open question, but for now it's capital with a built-in customer relationship attached, which is a structurally different kind of round than most Western AI raises.

The Platform Bet: Software Brain, Not Robot Body

PsiBot doesn't build robots. It builds vision-language-action (VLA) models and dexterous-manipulation algorithms and licenses them to other robot makers — a platform strategy that looks more like an operating system than a hardware OEM. That's a deliberate contrast with companies like Unitree Robotics and AgiBot, which build and sell physical humanoids directly. PsiBot is betting that as the number of robot-body manufacturers proliferates and commoditizes, the actual moat sits one layer up, in whichever software brain is generalizable enough to run across the widest range of hardware.

It's the same wager Nvidia has effectively made with its own robotics foundation-model work, and the same wager driving valuations across the physical AI opportunity more broadly: if the software layer generalizes, it captures value across every downstream hardware maker that licenses it, the way Android captured value across phone OEMs that never controlled the operating system themselves.

China's Embodied AI Unicorn Count: 3 to 22 in One Year

At the end of 2025, exactly three Chinese embodied AI companies — Unitree Robotics, AgiBot, and Galbot — had crossed the CNY 10 billion (roughly $1.4 billion) valuation threshold. By the first half of 2026, that list had grown to at least 22, PsiBot among the newest additions. Sector-wide, Chinese robotics startups raised roughly $15.5 billion in the first half of 2026, a fivefold jump from the same period a year earlier, spread across 126-plus disclosed deals in Q1 alone.

Figures from YICAI/Shanghai Stock Exchange reporting, BigGo Finance, Bloomberg, and TechTimes as of July 2026.

PsiBot's Funding History and the Investor Mix

PsiBot's cap table tells its own story about who's funding China's embodied AI buildout. The earliest angel round drew almost entirely state-linked and industrial capital rather than traditional venture funds — Guokai Finance, Guozhong Capital, and a fund affiliated with state broadcaster CCTV. The Pre-A round brought in Shanghai-based Xuhui Capital alongside other local state funds. The current round swaps in strategic industrial money — an automaker and a components supplier — rather than a marquee global VC name leading the check.

That mix is a useful proxy for how Beijing is directing capital into the sector: less reliance on Silicon Valley-style venture funds, more direct participation from state-linked funds and industrial strategics with a use case already lined up. It's capital formation built for deployment speed, not just financial return — which tracks with how fast the unicorn count moved this year.

What This Means for the Global Robotics Race

The US humanoid race — Figure, Apptronik, 1X, Tesla Optimus — has largely been a hardware story, with each company building both the robot and the model that runs it, as covered in our humanoid robotics comparison. PsiBot's platform approach — and China's broader push toward a software layer decoupled from any single hardware maker — is a different bet on where the category's value concentrates. If VLA models turn out to be the scarce, generalizable resource, a company that licenses across dozens of Chinese robot OEMs could end up better positioned than any single vertically-integrated hardware player, Chinese or American.

It also raises the stakes on chip access. Training and running large-scale VLA models at the pace China's embodied AI sector is now funding requires serious compute, the same compute constrained by ongoing export controls — a dynamic we've tracked in detail on the AI chip export controls front. A software-platform strategy like PsiBot's is, in part, a hedge against hardware-supply uncertainty: license the brain broadly enough, and you diversify your exposure to any single chip or robot-body bottleneck.

The Bottom Line

A $1.48 billion valuation for a two-year-old company would be startling news in most sectors; in China's embodied AI market in 2026, it makes PsiBot the 22nd name on a list that had three entries seven months ago. The more durable signal isn't the number — it's that an automaker chose to lead the round instead of simply buying the technology later, and that PsiBot chose to sell software instead of hardware. Whether that platform bet pays off depends on whether VLA models actually generalize across robot bodies the way operating systems generalized across devices — the same open question every embodied AI investor, in China or anywhere else, is trying to answer right now.