Humanoid Robotics Funding: $11B Boom Hides Deep Concentration

A record year built on a handful of checks

Humanoid robotics is having the kind of year that generates breathless charts. Venture investment into humanoid startups has reached $11 billion in 2026 so far, more than double the full-year record set in 2025 1. The biggest rounds have gone to a small cluster of names, including Neura Robotics, XPeng Robotics, AI2 Robotics, Sharpa and Apptronik 1.

The headline total is real. But it does not describe a broad-based boom in robotics investment. Across all three bodies of data, the same pattern appears: capital is pooling in very large rounds at the top while the rest of the market moves sideways or shrinks. Humanoids are the sharpest version of a trend running through robotics generally.

Strip out the megadeals and the boom fades

The clearest evidence comes from looking below the largest rounds. Robotics funding in rounds under $50 million totals $882 million so far in 2026, down from $1.02 billion over the same stretch last year 2. Put simply, the smaller-check end of the market is contracting even as aggregate numbers climb.

A longer view points the same way. An F-Prime analysis built on PitchBook data finds that the recent surge is concentrated in rounds above $50 million, while early- and mid-stage robotics funding has been essentially flat for five years 3. That matters because it frames the current spike as a top-heavy event, not a rising tide.

Broad category labels can also inflate the sense of scale. Crunchbase counts $47.4 billion in "physical AI" funding for the first half of 2026, but that bucket includes Waymo, defense and aerospace alongside robotics 3. Comparing that figure with robotics-only or humanoid-only totals mixes very different businesses, and it overstates how widely the money is spreading 3.

Early stage: thinner share, not a freeze

None of this means young robotics companies have been shut out. Seed and Series A rounds still account for 54.3% of 2026 robotics deals, although that is down from 66.0% in the same period of 2025 2. First financings actually rose to 26 2. The deal count suggests investors are still placing early bets. What has changed is the share of the market those bets represent.

The more telling shift is in round size. Median seed rounds sit at $3.52 million, Series A at $18 million, Series B at $50 million, and Series C at $104.9 million 2. That steep climb means a startup that graduates to later stages enters a different capital universe. A company that stalls between rounds faces a market where the gap between a first check and a follow-on has become enormous 2.

One way to read this is that investors have chosen their likely winners and are concentrating firepower on them. Humanoid makers fit that profile well. Building a general-purpose robot body, along with the foundation models meant to control it, needs heavy hardware spending and long development cycles, which naturally favors very large rounds.

Valuations priced like software

The concentration has also produced aggressive pricing in adjacent categories, especially companies that supply training data for robot models. XDOF is reportedly in talks for a $1.2 billion Series B, roughly 24 times its approximately $50 million in annualized revenue 3. That is a multiple more typical of software. Yet the underlying business depends on sensors, operators and headcount, and it is unclear whether its economics can hold up once the current funding rush cools 3. The same yardstick offers a check on how investors are valuing other robot-data plays 3.

This is where the robot foundation model story and the funding story meet. Better models need large volumes of real-world physical data, and investors are paying up for anyone who can produce it at scale. Whether that data work turns into a durable, defensible business or remains a labor-heavy service propped up by abundant capital is still an open question.

The takeaway

Read together, these figures describe a market that is deep at the very top and shallow underneath. The $11 billion humanoid figure 1 is best read as a signal of conviction in a few well-funded platforms, not as evidence that robotics as a whole is flush. Smaller rounds are shrinking 2, mid-stage funding has stagnated for years 3, and some valuations depend on assumptions about data businesses that remain untested 3.

For founders, this suggests the hardest point in the journey may be the jump from Series A to the later stages, where median round sizes rise sharply. For investors, the risk is familiar. When capital crowds into a small number of bets, outcomes become binary, and today's record totals could look very different if even a few of those flagship companies stumble.