Crunchbase Data Shows Q3 Robotics Startup Funding Jumps 185% As US And China Lead Physical AI Push

Venture investments in robotics startups have surged throughout 2026, with the majority of capital directed toward businesses located in the U.S. and China, according to Crunchbase data. Globally, venture funding for robotics startups reached $9.9 billion in Q3, significantly surpassing the $4.3 billion recorded in Q3 2025, as reported by Crunchbase 1.

This follows a consecutive run of quarters bringing in roughly $10 billion or more. Taken together, robotics sector funding between Q1 and Q3 2026 has increased by approximately 185% year over year.

China and the US lead

China has captured the largest share of robotics funding this year with $13.2 billion invested, closely followed by the U.S. at $12.1 billion. Germany holds the third-position for startup funding in this sector, trailing significantly with $1.7 billion.

Among the major financing rounds in China, Guangzhou-based Xpeng Robotics — a creator of physical robots and AI models — secured $900 million, while Shenzhen-based AI² Robotics closed a $736 million Series C round to develop intelligent robots geared toward manufacturing, healthcare, and services.

Domestically, U.S. companies commanding billion-dollar rounds this year include Travis Kalanick’s Los Angeles-based Atoms (incorporating its ghost kitchen enterprise alongside transport and mining robotics divisions) and Pittsburgh-based Skild AI, which secured $1.4 billion for its robotics foundation model, Skild Brain.

Germany also saw a major financing event with Neura Robotics pulling in $1.4 billion to fund its cognitive robots, along with the data and AI infrastructure needed to generate intelligence.

Building out the data and model layer

“We’re just seeing physical AI take off,” Ulrik Stig Hansen, co-CEO of San Francisco-based Encord — a physical AI data layer provider — shared with Crunchbase News during the inaugural Long Horizon Physical AI Summit held in San Francisco earlier this month.

While compute and storage infrastructure are expanding to support these technologies, development in the physical realm demands time, particularly when contrasted with large language models trained on vast amounts of internet text. Hansen explained, “Physical AI is very different because there is no internet of physical AI data, and so the data first has to be collected, and you have to convert it from sensor tokens to action tokens.”

Because physical AI datasets are magnitudes larger and demand multimodal data merging force, torque, tactile, and vision inputs, Hansen added that “storing all that data, curating all that data, enriching all that data, and validating that all the data works for these systems is a very expensive operation.”

The industry remains anchored in the pre-training stage of physical AI, though it will soon transition into the post-training phase, he observed.

More winners

Nearly 75 robotics enterprises have achieved unicorn status after raising capital over the past three years. More than half of them — specifically 45 — attained this ranking just this year. Among the top-valued robotics startups, numerous competitors are designing competing models and data structures alongside diverse vertical use cases spanning logistics, heavy industry, manufacturing, and household applications.

Even though funding is skyrocketing and new unicorns continue to emerge, the long-term trajectory of the industry stays open as companies race toward upcoming breakthroughs in the coming year.

“There are a lot of companies that are raising capital, and I think a lot of them are going to win, just because there’s so many different types of applications to the physical world,” Hansen stated.

?Frequently Asked Questions

01What caused the surge in robotics funding in 2026?

The surge is primarily driven by advancements and investments in physical AI, with massive backing from the U.S. and China.

02Which countries lead in robotics startup funding?

China and the U.S. lead the sector by a wide margin, followed by Germany in third place.

03What makes training physical AI different from large language models?

Unlike text-based large language models, physical AI lacks a pre-existing “internet of physical data,” requiring developers to collect, convert, and manage massive multimodal datasets involving vision, tactile, force, and torque data.

Related Crunchbase queries:

• Global Physical AI Venture Funding In 2026
• Global Robotics Venture Funding In 2026

Related reading:

• Sector Snapshot: Robotics Startups On Fire As Venture Funding Surges To Record Numbers In 2026
• Embodied AI Fuels Record Robotics Funding In China As IPO Momentum Builds
• VCs Pour Billions Into Physical AI As The Next Wave Of AI Investing Takes Shape

Illustration: Dom Guzman