X Square Robot Crosses $2.8B as China's Embodied-AI Race Heats
A Shenzhen startup building foundation models for robots closed a Series C backed by China's biggest tech names — and the valuation says the physical-AI bet is going long.
X Square Robot, a Shenzhen startup building foundation models for physical robots, has closed a Series C at a valuation above $2.8 billion — the culmination of four consecutive financing rounds and a roster of backers that reads like a directory of Chinese tech's most powerful checkbooks. IDG led the latest round, joining prior investors that include HongShan, Xiaomi, Meituan, Alibaba, and ByteDance. The concentration of strategic capital around a single embodied-AI company is a statement about where China thinks the next platform layer sits: not in the model that writes text, but in the model that moves a body through the world.
The Foundation-Model Layer for Robots
X Square Robot's bet is that robotics is following the same arc as language AI. The last generation of robots was programmed task by task — bolt this, sort that, follow this fixed path. The thesis behind embodied-AI foundation models is that a single large model, trained on enough physical interaction data, can generalize across tasks the way a language model generalizes across prompts. Instead of hand-coding a robot to fold a shirt, you train a model that has learned the physics of manipulation broadly enough to fold a shirt it has never seen, in a kitchen it has never entered.
That generalization is the whole game. A robot that needs bespoke programming for every new task is a piece of industrial equipment; a robot running a foundation model that adapts to novel situations is a platform. The companies that own the model layer — the "brain" that any robot body can run — capture the position that Nvidia holds in compute and that OpenAI and Anthropic hold in language. X Square is racing for that seat in the physical world, and its investors are pricing the possibility that it wins.
Why Everyone Wrote a Check
The investor list is the story as much as the valuation. Xiaomi builds consumer hardware and is pushing into its own robotics and manufacturing automation. Meituan runs one of the world's largest delivery-logistics operations, where autonomous physical systems translate directly into margin. Alibaba and ByteDance bring cloud infrastructure, data scale, and platform reach. HongShan — the firm formerly aligned with Sequoia's China business — anchors the financial syndicate, and IDG leads the newest round. When strategics this varied all take positions in the same company, they are hedging against the emergence of a dominant embodied-AI platform they don't control. Owning a piece of the likely winner is cheaper than being locked out of it.
There is also a national dimension. China has made physical AI and humanoid robotics an explicit industrial priority, framing embodied intelligence as a strategic technology on par with semiconductors and AI models. A domestic foundation-model company for robots, funded by the country's leading platforms, fits directly into that agenda — a homegrown brain for a homegrown fleet of machines, insulated from the export controls and supply-chain leverage that constrain China's access to frontier compute.
From Demo to Deployment
The hard part is the same one facing every embodied-AI company, in China or the U.S.: the gap between an impressive demo and a robot that earns its keep in a warehouse, a factory, or a home. Manipulation is unforgiving. A language model that produces a slightly wrong sentence is a nuisance; a robot that grips a fragile object slightly wrong breaks it. Getting from "generalizes in the lab" to "reliable enough to deploy at scale" has humbled well-funded teams before, and a $2.8 billion valuation is a bet on execution that hasn't been fully proven yet.
What X Square has that a Western counterpart lacks is a uniquely dense deployment environment. China's manufacturing base, logistics networks, and hardware supply chains give an embodied-AI company an enormous surface of real-world tasks to train on and sell into — and strategic investors like Meituan and Xiaomi can hand it deployment channels most startups would spend years building. The data flywheel that matters for physical AI — robots doing real work, generating real interaction data, improving the model — is easier to spin up when your cap table includes the companies that operate the factories and the delivery fleets.
The Physical-AI Money Is Getting Serious
X Square's raise is one more data point in a year where capital has rotated hard toward physical AI. Humanoids are moving from pilots to production lines, warehouse robots are scaling, and the foundation-model layer beneath all of it is where the largest valuations are forming. A $2.8 billion price on a company whose core product is a robot brain reflects a belief that the value in robotics is migrating up the stack — away from the actuators and toward the intelligence that controls them.
Whether X Square becomes the Android of robots or a cautionary tale about the distance between generalization and reliability, the round marks how far the physical-AI thesis has traveled. A few years ago, a robot company raised money on the strength of its hardware. Today, one raises $2.8 billion on the strength of a model — and China's biggest platforms are all betting it's the one they'll want to own.