Robotics startups raised more money in six months than in any full year before 2025.
Globally, the sector pulled in $18.8 billion in the first half of 2026, according to Crunchbase data. That figure eclipses the $15 billion raised in all of 2025 and blows past the previous peak of $14.1 billion in 2021. This is not incremental growth. It is a structural reallocation of capital away from software-only bets and into companies that can put AI into motion.
The numbers are staggering. The question is whether they are sustainable.
The Billion-Dollar Club Gets Crowded
The surge is concentrated in embodied AI and humanoids. Three deals define the scale of the bet.
Saronic, a defense robotics startup, raised a $1.75 billion Series D in March, led by Kleiner Perkins, setting its valuation at $9.25 billion. In January, Skild AI closed a $1.4 billion Series C led by SoftBank, tripling its valuation to over $14 billion. Germany's Neura Robotics secured up to $1.4 billion in Series C funding from a consortium including Tether Holdings, Nvidia, Amazon, Qualcomm, Robert Bosch, and Schaeffler.
These are not software multiples. These are industrial-scale capital commitments that assume a company will eventually manufacture complex physical systems at volume and at margin.
China is matching the intensity. Shihang Intelligent raised a $1 billion Series A. Unitree Robotics filed for a Shanghai listing targeting a $3 to $7 billion valuation. Robotphoenix debuted on the HKEX, soaring nearly 80 percent before settling lower. The US-China robotics race is now a capital arms race, and both sides are fully funded.
M&A activity reinforces the thesis. Meta bought San Diego's Assured Robot Intelligence. Symbotic acquired Austin's Fox Robotics. The large platforms are not waiting for the startups to mature; they are absorbing the talent and IP now.
Why Capital Stopped Fearing Hardware
For a decade, venture capital treated robotics as a graveyard. Long development cycles, brutal unit economics, and integration nightmares made software look like the only rational bet. That logic has inverted.
The reasoning boom created a new problem: foundation models that can think but cannot act. Investors now see the physical world as the next compute surface. Hardware is no longer a risky side bet. It is the inevitable extension of the same AI thesis that drove the last five years of software returns.
The concentration of capital tells the story. Rohit Yadav notes that 70 to 80 percent of robotics dollars are going into rounds above $100 to $200 million. This is a winner-take-most dynamic. A handful of companies are absorbing the vast majority of available capital, talent, and strategic attention.
One likely consequence: smaller robotics startups, including profitable niche players in industrial automation, will struggle to raise follow-on rounds. When the sector's capital is locked up in a dozen mega-rounds, the feeder system of early-stage companies gets starved. That is not reported fact yet — but it is the logical endpoint of the concentration pattern.
Legacy automation companies face a different threat. Capital that once flowed to proven industrial robotics firms is being redirected toward companies that promise general-purpose intelligence in a physical form factor. The reallocation is happening inside 18 months.
The Correction Is Already Priced In
Here is the structural risk that the consensus is missing.
The humanoid robotics sector raised a record $13.8 billion globally in 2025, and 2026 is on track to exceed it. Goldman Sachs projects the humanoid robotics market to reach $38 billion by 2035. Those projections are driving valuations. They are not driving revenue.
In 12 to 24 months, at least three of the mega-round startups — likely Saronic, Skild AI, and Neura Robotics — will face severe production delays or unit economic failures. The mechanism is straightforward. Scaling hardware is not like scaling software. Supply chains have hard constraints. Manufacturing tolerances at volume are unforgiving. Regulatory approvals for autonomous physical systems, especially in defense and human-adjacent environments, are slow and political. The gap between a polished demo reel and a deployed fleet with positive gross margins is measured in years, not quarters.
When those failures arrive, the correction will be sharp. I expect a 30 to 40 percent contraction in robotics venture funding as investors shift from funding hype to demanding deployed fleet metrics and positive unit economics. The same funds that wrote billion-dollar checks on the promise of embodied AI will freeze when the first production milestones are missed.
The transmission mechanism is systemic, not isolated. When 70 to 80 percent of all robotics funding sits in a dozen companies, a correction in those specific names does not just hurt their investors. It freezes the entire ecosystem. Here is how it cascades: a missed production milestone at a flagship startup triggers a valuation markdown. The markdown spooks limited partners, who pause new commitments. Venture firms, facing a suddenly illiquid portfolio, stop writing checks to earlier-stage companies to preserve dry powder. The halo effect that helped smaller startups raise evaporates. Talent that joined the sector for equity upside leaves. The cycle turns fast.
The washout is not a bug. It is a necessary clearing mechanism. Robotics as a sector needs the discipline that comes from capital scarcity. The companies that survive a 30 to 40 percent funding contraction will be the ones that figured out how to build, ship, and service physical systems at margin. The ones that don't will be acquired for pennies on the dollar or shuttered entirely.
What To Watch When the Music Stops
Operators and investors should track three signals.
Production milestones. Which companies ship working units to paying customers, not pilot partners, in 2027. A demo is not a deployment.
Gross margin disclosures. Any startup raising at a multi-billion-dollar valuation and refusing to disclose unit economics is hiding something. Demand the numbers.
Fleet size. The companies that survive the correction will be the ones with hundreds or thousands of units in the field generating data that improves the product. A company with ten units and a $14 billion valuation is a financial instrument, not a robotics business.
For corporate acquirers, the coming correction is an opportunity. Meta and Symbotic moved early. Others should prepare balance sheets now to absorb distressed assets when the funding window narrows.
The Hardest Part Starts Now
The $18.8 billion is both a milestone and a warning. The capital is there. The thesis is funded. The hard work of delivering physical systems at scale begins now, and it will separate the real from the hype inside 18 months.
The correction will be brutal. It will also be the best thing that ever happened to the robotics industry. Capital discipline forces focus. Focus forces execution. Execution is the only thing that builds a company that lasts.
