AI² Robotics Bets $735M That Wheels Beat Legs
The Shenzhen startup raised three-quarters of a billion dollars at a ~$2.8B valuation for a humanoid that rolls instead of walks — a deliberate wager that the fastest path to useful robots skips the hardest problem in the field.
AI² Robotics raised roughly $735 million in a new round that pushes its valuation past 50 billion RMB — about $2.8 billion, with some accounts putting it closer to $3 billion. The number alone would make it one of the largest humanoid raises of the year. But the more interesting thing about the Shenzhen company is what it decided not to build. Its flagship, AlphaBot, is a humanoid from the waist up and a wheeled platform from the waist down. AI² looked at bipedal locomotion — the marquee challenge the entire humanoid field has organized itself around — and chose to route around it.
The case against legs
Bipedal walking is the hardest, most fragile, most expensive problem in humanoid robotics. It burns compute and battery on balance, breaks in unpredictable ways, and introduces a safety surface — a 60-kilogram machine that can topple — that terrifies every factory manager who might otherwise deploy one. It's also, for most real jobs, not the point. Warehouses, production lines, and logistics floors are flat. The work is in the hands and the torso: perceiving, reaching, grasping, manipulating, moving objects from A to B.
AI² built AlphaBot around that observation. The company says it chose a wheeled base with a humanoid upper body specifically to guarantee industrial-grade stability, speed, and safety — trading the mobility range of legs for mechanical simplicity and durability. A wheeled robot doesn't fall over. It doesn't spend its power budget staying upright. It's cheaper to build, easier to certify, and far more predictable on a factory floor. What it gives up — stairs, rough terrain, the full human envelope of movement — mostly doesn't exist in the environments where these robots will actually earn their keep first.
The intelligence lives up top. AlphaBot is driven by AI²'s "Alpha Brain" foundation model, which the company says can interpret complex commands, manipulate objects, and carry out goal-based tasks rather than pre-scripted routines. That's the embodied-AI thesis in miniature: put the hard problem in the model, not the mechanics.
Who's paying for it
The investor list is where this stops being a pure-tech story and becomes a read on Chinese industrial policy. AI²'s round drew from three distinct pools at once. State capital, including the National Small and Medium Enterprises Development Fund. Industrial corporates with real deployment surfaces — Sino Biopharmaceutical and, notably, the Moutai Group, the liquor giant, signaling interest from manufacturing sectors far outside tech. And financial firms like CICC Capital and GSR Ventures.
That combination — government funds, strategic industrials, and institutional finance in a single round — is the signature of a technology Beijing has decided is strategically important. It's not venture money chasing a demo. It's a coordinated bet that embodied AI is national infrastructure, funded by the entities that would both build it and buy it.
A wave, not a one-off
AI²'s raise didn't happen in isolation. In roughly the same window, fellow Chinese robotics outfit X Square Robots secured funding at a comparable ~$2.8 billion valuation, part of a broader surge that has made humanoids one of the most heavily capitalized sectors on the planet. The scale is now hard to overstate: robotics startups have raised $18.8 billion globally in 2026 so far — already past the roughly $15 billion invested across all of 2025, with months left on the calendar.
Inside that flood, two design philosophies are hardening into camps. On one side, the bipedal maximalists — Figure, Tesla's Optimus, the Unitree and UBTech humanoids — betting that a general-purpose robot has to be shaped exactly like a person to slot into a world built for people. On the other, the pragmatists like AI², betting that the fastest route to robots that pay for themselves is to solve the valuable 80% now and leave legged locomotion for later. AlphaBot is the clearest, best-funded statement yet of the second camp.
Why the wheeled bet might be right
The pragmatist case has history on its side. Warehouse automation didn't wait for humanoids — it shipped wheeled AMRs and robotic arms that did narrow jobs profitably for a decade. AI²'s wager is that the winning near-term form factor is a fusion: the dexterity and general intelligence of a humanoid upper body, mounted on the boring, reliable, already-solved wheeled base that industry already trusts. It sidesteps the physics problem that keeps bipedal robots in pilot programs and gets a capable manipulator onto a real floor faster.
The risk is that it's a local maximum. If bipedal locomotion gets solved — cheaply, reliably, safely — the wheeled compromise could look like a transitional hack, useful until the "real" humanoids arrive. AI² is effectively betting that "later" is far enough away that a wheeled platform earns its keep and its data advantage in the meantime, and that the manipulation intelligence in Alpha Brain transfers to whatever body comes next.
At $735 million and a near-$3 billion valuation, the market is willing to fund that bet at serious scale. The humanoid field spent years insisting the robot had to walk. AI² just raised three-quarters of a billion dollars arguing that, for the jobs that matter first, it doesn't.