Recommended Articles
Fundraising

Humanoid Robots Have Raised $11.8 Billion. Their Revenue Reckoning Is Next.

By Angeled · October 27, 2026 · Fundraising
Humanoid Robots Have Raised $11.8 Billion. Their Revenue Reckoning Is Next.
Humanoid Robots Have Raised $11.8 Billion. Their Revenue Reckoning Is Next.

Something strange is happening in humanoid robotics. The money has arrived — in extraordinary quantities — but the revenue has not. A funding tracker current to October 2026 counted about $11.8 billion raised across 26 tracked humanoid companies, before counting internally funded programs like Tesla's and Xiaomi's, which it puts separately at more than $30 billion. Yet the companies furthest along report revenues that would embarrass a mid-sized software startup.

UBTech, the Chinese company furthest along commercially, reported about $86.9 million in revenue for the first half of 2026. Agility Robotics reported $1.8 million in net sales for all of 2025. Figure AI and Apptronik, two of the largest private names, disclose no revenue at all. Figure alone has drawn roughly $1.9 billion and was valued at $39 billion in its September 2025 Series C. The gap between capital raised and cash returned is the defining fact of this sector — and the next twelve months will test whether it closes.

October's Deals Show the Bet Is Still Growing

If anything, October 2026 accelerated the bet rather than pausing it. On October 7, Mecka AI raised a $60 million Series B led by Sequoia, with Nvidia, Microsoft's venture fund M12, Qualcomm Ventures and Samsung participating, to collect and analyse human motion data for training humanoid robots. Mecka pays people to record themselves doing everyday tasks — making coffee, fixing cars — while wearing body sensors, building the datasets that robot models learn from. The company was reportedly nearing the round at a $500 million valuation, though it did not confirm the figure.

On October 19, Dream Labs announced $90 million in seed funding led by B Capital and Kindred Ventures for Dream-1, a robot foundation model that turns camera feeds and plain-English instructions into motor commands, sold as an API rather than as a robot. On October 26, Beijing-based CASBOT announced a Series A++ round of several hundred million yuan for its humanoid and industrial robotic systems, reporting more than RMB 400 million in orders connected to footwear manufacturing — company-reported orders, not recognised revenue, but a more specific commercial application than most general-purpose demos offer. And on October 20, YITA, a Chinese robot hardware startup founded only in May 2026, raised more than RMB 100 million across seed rounds for force-controlled arms and joint modules.

The pattern is clear: investors are funding every layer at once — data, models, components, and full robots. That breadth is a strength, but it also means no single layer can hide behind the others' progress anymore.

Why Demos No Longer Suffice

The humanoid pitch has always rested on generality: one machine that can do many jobs, learning from data, improving with scale. The problem is that generality is expensive to prove and cheap to claim. A video of a robot folding laundry is not evidence of a business; it is evidence of a laboratory. Enterprise buyers evaluate cycle times, uptime, servicing requirements, and return on automation — the unglamorous metrics that decide whether a pilot becomes a purchase order.

Sam Altman's reported comment this month that OpenAI will build a humanoid of its own, with a prototype as soon as 2027, sharpens the point. When the company most startups treated as a supplier of robot brains becomes a competitor in bodies, the moat question changes. Data, deployment experience, and manufacturing capability matter more than model cleverness alone. Mecka's bet — that human motion data is the scarce input, just as labelled text was for language models — is one answer. Dream Labs' bet — that software across many hardware makers beats one vertically integrated robot — is another. Both will be judged on the same thing: robots working reliably, repeatedly, in paying customers' facilities.

What It Means for Founders and Investors

For founders in embodied AI, the fundraising bar is moving from "show the robot" to "show the deployment." The metrics that will win the next round are units shipped, hours operated, failure rates, and repeat orders — not video views. Founders should also choose their layer honestly: building data, models, components, or full systems are different businesses with different capital needs, and the market is now funding all four, which means each will be compared against specialists, not generalists.

For investors, the discipline is to separate the technology risk from the commercial risk. The technology is plainly advancing. The commercial question — who pays, for what task, at what price, and how soon — is almost entirely unanswered at the scale of the capital deployed. The companies that answer it first, even in a narrow niche like CASBOT's footwear manufacturing, will command the next funding cycle. The rest will discover that $11.8 billion of capital is not a moat; it is a clock.

Angeled.com connects founders with investors.

Source: ainvest.com
AInvest, "Bloomberg: Is that a robot?" (October 2026)
Source: techcrunch.com
TechCrunch, "Robot data startup Mecka AI nabs $60M from Sequoia" (October 7, 2026)
Source: techstartups.com
Tech Startups, "Startup Funding News Today, October 26, 2026: Ailytics, Approaching.AI, CASBOT, Volare & More" (October 26, 2026)
Source: trendintech.com
TrendInTech, "Dream Labs Scores $90M for Robot AI Model" (October 2026)
Source: technode.com
TechNode, "Robot hardware startup YITA secures funding from Galbot and Fourier" (October 21, 2026)

Read Next