Sometime around the middle of this year, robotics stopped being a hardware bet and started being an AI bet. The capital followed the reclassification, and it did not move gradually.
Robotics startups have raised roughly $18.8 billion globally in 2026, according to Crunchbase data. The full-year 2025 total was $15 billion. The previous record, set at the top of the last cycle in 2021, was $14.1 billion. Both marks fell before summer. There is still a quarter of the year left.
That is not a trend line bending upward. That is a sector being repriced.
Robotics spent a decade carrying a specific reputation among venture investors: capital-hungry, slow to margin, a business where the atoms eat the returns. What changed is not the hardware. It is that vision-language-action models made general-purpose machines look like a software problem with a chassis attached — and software problems are what this generation of venture capital knows how to underwrite at scale.
Where the Money Went
The largest disclosed round of the year is not a humanoid. In March, Austin-based Saronic, which builds autonomous surface vessels for defense customers, raised $1.75 billion in a Series D led by Kleiner Perkins at a $9.25 billion valuation — more than double where its Series C landed in 2025.
The rounds behind it map the thesis. In January, Pittsburgh-based Skild AI raised close to $1.4 billion led by SoftBank Group, with Nvidia's NVentures, Bezos Expeditions, Samsung, LG and Salesforce Ventures participating. The round valued the company above $14 billion, roughly triple its valuation seven months earlier. Skild is building what it calls an omni-bodied foundation model — one brain, any body.
“The Skild Brain can control robots it has never trained on, adapting in real time to extreme changes in form or environments. The model is forced to adapt rather than memorize — much like intelligence in nature,” said Deepak Pathak, co-founder and CEO of Skild AI.
In June, Germany's Neura Robotics announced a Series C of up to $1.4 billion led by Tether, with Amazon, Nvidia, Qualcomm, Bosch and the European Investment Bank joining — the largest robotics round ever raised in Europe. Beijing's Shihang Intelligent took $1 billion in a Series A. Apptronik pulled a $520 million extension onto its Series A in February, bringing that single round past $935 million, with John Deere joining Google and Mercedes-Benz.
Rivian spinout Mind Robotics raised twice in three months: $500 million in March co-led by Accel and Andreessen Horowitz, then $400 million in May led by Kleiner Perkins, which led two of the year's five biggest deals. The investor list is the real signal: alongside the usual growth funds sit Nvidia, Amazon, Qualcomm, Bosch, John Deere, a stablecoin issuer and a national development bank. Strategic and industrial money, not just financial.
Why This Matters
The first thing the numbers describe is a rotation. Excluding the outlier rounds at OpenAI and Anthropic, hardware companies now account for nearly one-third of all US venture capital invested this year, per PitchBook data analyzed by Silicon Valley Bank. One in five global investors now allocates at least 10% of its deals to US hardware — a share that has more than doubled since 2022. SVB projects $120 billion into hardware companies by year-end if the pace holds. Capital that spent three years chasing application-layer software has decided the durable moat is physical.
The second thing is that almost none of this is underwritten by revenue. Skild went from zero to about $30 million in revenue over a few months in 2025 — genuinely fast, and roughly 470 times below its valuation. That makes it one of the cohort's leaders. Meanwhile the demand side looks thin: in a Prologis survey of 400 warehouse operators, one in three large companies had automated fewer than 25% of their warehouses, and fewer than half of facilities over 250,000 square feet use AI in daily operations. The stated barriers were high upfront costs and long payback periods — the exact objection the funding boom has not yet answered.
The third is geography, the least appreciated of the three. China-based robotics companies raised $5.6 billion across 176 deals through mid-May, already past the $4.3 billion they raised in all of 2025. China now accounts for more than 43% of global robotics venture investment. That money behaves differently: more deals, smaller median checks, and extraordinarily fast follow-ons. Spirit AI raised $290 million in February, then $145 million more in April; Galaxea AI ran the same play at similar size.
China also has something the US robotics sector does not: exits. Robotphoenix listed on the HKEX in May and closed its first full day up nearly 80%, though shares have since faded. Unitree filed in March for a Shanghai listing targeting a $3 billion to $7 billion valuation. American robotics companies are raising at higher valuations into a quieter IPO window, and taking liquidity through M&A instead: Symbotic bought Fox Robotics, Skild acquired Zebra Technologies' robotics arm, and Meta absorbed Assured Robot Intelligence.
“Solving intelligence for the physical world unlocks enormous commercial value and long-term strategic national importance,” said Rita Waite, a partner at IQT, the CIA-affiliated venture firm, explaining its Skild investment. That sentence is doing a lot of work. When national-security framing enters a funding thesis, price discipline usually exits.
What to Watch
Whether Unitree prices, and where. A successful Shanghai listing sets the first real public comparable for embodied AI, and would pull others out behind it.
Whether the second half sustains the first. The $18.8 billion figure reflects data through late June; September deals like Lyte's $165 million Series C at a $1.6 billion valuation suggest the pace has not broken.
Whether warehouse automation adoption moves. It is the cleanest available proxy for whether any of this converts. If the Prologis numbers look the same next year, the capital arrived years ahead of the customers.
“Solving intelligence for the physical world unlocks enormous commercial value and long-term strategic national importance.”— Rita Waite, Partner, IQT