Is Physical AI Heating Up or Cooling Down: Financing Cools While Tech Giants Enter

Is Physical AI heating up or "cooling down"? Science and Technology Innovation Board Daily November 03, 2026 12:35:50 Updated November 03, 2026 13:19:44

From the past "flexing athletic muscles" to today's "flexing brains," among this year's hottest buzzwords in the AI world, Physical AI certainly occupies a place. Looking across this year's robot exhibitions, concepts centered on the "brain" such as "world models," "one brain, multiple forms," and "eyes-brain-hands" have been pushed to the forefront by major companies as core selling points.

At the World Physical AI Conference in October, Unitree Technology founder Wang Xingxing said that with the release of GPT-6, Physical AI has now shown signs of being broken through, and a key breakthrough is likely to come in the next 1–2 years or 2–3 years. Currently, the core challenge for embodied intelligence to pass the ChatGPT moment is the alignment between the input and output of AI models and real robots.

As "can it be deployed" and "can it do work" become the focus of the industry, related concepts around "Physical AI" carry a great deal of industry expectation, and the world model is one of them. As the "brain" for robots to understand the physical world and interact with the external environment, the world model is held to high expectations.

But the other side of high expectations is the possibility of falling short. Since the second half of the year, several "brain"-related companies have still secured large single-round financings, but there is more and more news of cooling in the entire embodied intelligence primary market, and the risks, bubbles, and anxiety surrounding Physical AI have followed.

Is Physical AI actually heating up, or is it "cooling down"? The reporter found that beneath this boom, primary-market capital, startups, tech giants, and the A-share industrial chain are showing markedly different temperatures.

▌2026 is the year of the "brain": hot money flows toward "Physical AI"

Data from Cailian Press Venture Capital Connect shows that, according to incomplete statistics, before 2025, primary-market financing events in Physical AI/world models were mostly in the single digits, rising to 23 in 2025. As of now, 2026 has already reached more than 100, and multiple financing rounds by a single company have become the norm.

Judging from the distribution of financing timing, June to September 2026 was the period when the track "absorbed large amounts of money" most intensively. For example, LiberAI focuses on physical world models and completed four financing rounds from May to September, totaling nearly 1 billion yuan; Liu Songming is regarded as having the "most top-tier" founder profile among the "little geniuses" currently pursued by VCs; Manifold Space raised from a seed round in July 2025 all the way to a Pre-A+ round in June 2026, completing six financing rounds in total over the past year, with Pre-A round financing totaling nearly 1 billion yuan.

In the world model field, it is not uncommon for startups to complete several financing "leaps" in a short period. Although seed rounds, angel rounds, Pre-A rounds, and other pre-Series A projects currently account for more than 70%, mostly early-stage projects, there are quite a few cases of single rounds exceeding 100 million yuan or even 1 billion yuan.

These large financings are concentrated in several types of companies: one type builds world models/embodied brain foundations, such as LiberAI, Octopus Dynamics, Inverse Matrix Technology, Manifold Space, Moshen Intelligence, and Westlake Digital Intelligence; another type builds data infrastructure and simulation engines, such as Guanglun Intelligence, Hive Interconnect, Feijie Kesi, and Songying Technology.

There is also a category of companies focused on combining brains with dexterous tactile software and hardware, or possessing core hardware mass-production barriers. For example, Lingchu Intelligence completed angel and Pre-A round financing in March 2026, officially announcing a total of 2 billion yuan received; in April, state-owned capital such as Guotou Xiandao and Jingxi Ruiling invested again; in July, it completed an A round of over 100 million US dollars. Another example is Paxini, which focuses on tactile sensors and dexterous hands; in March 2026 it secured a B round of over 1 billion yuan, in August it completed a 1 billion yuan strategic financing, and in September it received another several hundred million yuan in B++ round financing, with cumulative financing exceeding 4 billion yuan.

▌Second-half primary market: financing turns cold, LPs begin to ask about real scenarios

Data from Cailian Press Venture Capital Connect shows that, according to incomplete statistics, there were 5 primary-market financing events in Physical AI/world models in October alone, indeed a decrease compared with before. Entrepreneurs and investors active on the front lines have already felt the cooling ahead of time.

"Everyone fired their ammunition too fast," sighed Wang Sheng, angel partner at Inno Fund. He said that many top VCs emphasized in May and June that the market was too hot and that they needed to slow down. "But this is a train, and stepping on the brake won't stop it all at once. Inertia caused the number of projects closed in July to drop by about 20 to 30 percent, and August fell another 20 percent from July. From the outside it feels like September was a cliff, but the braking had actually already begun earlier."

"Generally speaking, after the tide goes out, you see who is swimming naked. There is a lot of money in the market, that's true; people may just feel the pace is a bit slower, but that doesn't mean there is essentially no money," said Yu Kaicheng, director of the Westlake World Model Research Center. In his view, the commercialization of new technology is a long process: "Suddenly a batch goes up, then it cools for a while, then suddenly an even bigger batch goes up—this is very normal." This year, Westlake Digital Intelligence Awomo, where Yu Kaicheng works, was established and launched the "Concept World Model." Recently, this startup has cumulatively obtained over 100 million yuan in seed and angel round financing, becoming one member of the world model startup wave.

Over the past three years, embodied intelligence has achieved a major leap in primary-market financing, with capital continuing to be released. In 2024, embodied intelligence financing was still 14.05 billion yuan; by 2025, total financing in China's embodied intelligence and robotics sectors reached 73.543 billion yuan, with 744 investment and financing events for the year. In the first half of 2026 alone, total financing in China's embodied intelligence track already reached 93.5 billion yuan, five times higher than the first half of 2025; the number of financing events reached 322, up 137% year on year.

Many investors also said that they have indeed become more cautious recently, with the pace slowing somewhat, and that investment focus has shifted from the imaginative potential of technology to investment risk and the ability to deliver on business results. However, there is still some lag before the slowdown in deal pace is reflected in the written financing data.

In the first half of the year, a joke circulated in investment circles: "Little genius + physical AI + top-journal papers = money wired directly." Now that the market is cooling, the "little genius" in the joke is being re-examined. But in conversations, some investors cautioned that "little genius" is not a derogatory term in the AI era, and being labeled as such does not mean a project is not a good project.

Fan Chenchao, deputy general manager of Zhejiang Venture Technology, believes that compared with the previous mobile internet era, when entrepreneurship was mostly based on new business models, this generation of AI entrepreneurship is mostly technology entrepreneurship. Moreover, "little genius" is only one point of attention in investment decisions and belongs to the "people" factor. Whether a startup ultimately succeeds depends on the combined result of many factors.

He summarized a profile formula for a foundation-model startup team: core founder's sensitivity to advanced technology routes × team leadership/organizational capability × talent density (industry appeal) + (capital + computing power + engineering capability, etc.). He believes that "if any one of the first few capabilities is a weak link, or is zero, the final result will be very poor, or even directly go to zero." Much of the "addition" part in parentheses can be solved with money.

At the capital roundtable of the founding conference of the Westlake World Model Research Center, Wang Sheng, a partner at Inno Fund, said bluntly that the biggest challenge for "little geniuses" comes from "future corporate governance and operations, as well as the ability to grasp business opportunities amid development and change—this is indeed what they need to learn."

▌Big tech rushes in: focus on foundations and ecosystems

If we shift our gaze to big tech, the situation is quite different. Google, Nvidia, AMD, ByteDance, Alibaba and others are moving quickly in layouts related to world models and physical AI. Compared with startups, big tech pays more attention to laying out foundations and ecosystems and to finding new points of integration between physical AI and existing businesses, rather than directly entering the embodied intelligence "brain" track.

First, they are competing for the next-generation interaction "entry point." Take Google's and ByteDance's world models as examples; the core is real-time interaction with the virtual world.

At the end of October, Google DeepMind officially announced its latest-generation world model Genie 4 and demonstrated an experiment in which it worked with the large model Gemini 4 Argon to simulate a humanoid robot performing tasks.

There have also been frequent reports about ByteDance's world model benchmarked against Genie. For example, in June, reports said Wu Yonghui set a goal for the world model to release at least one version benchmarked against Genie 3 by the end of the year; in September, ByteDance founder Zhang Yiming personally led the project, coordinating internal cross-departmental teams and investing computing power and AI R&D resources.

It is understood that the core of this R&D effort is to try to connect ByteDance's AI models, cloud computing, content platforms and Pico hardware, making the world model a new flywheel. In October, Tim, founder of Media Storm, had already demonstrated ByteDance's world model in advance; ByteDance has not yet officially released it.

Second, hardware giants have begun to "build computing platforms for the next generation of AI," with AMD and Nvidia as examples.

On September 28, AMD announced an all-stock acquisition of Fei-Fei Li's World Labs for about $8.2 billion. Fei-Fei Li will join AMD as executive vice president and chief scientist. This is also seen as a landmark event marking physical AI becoming a new battleground for computing power.

AMD Chair and CEO Lisa Su framed this as a long-term bet on the next decade: "Building computing platforms for the next generation of AI requires a deep understanding of how models will evolve. We can use that insight to develop hardware, software and systems."

As another major hardware chip company, Nvidia officially released in June this year the open world foundation model NVIDIA Cosmos 3 for physical AI. At the same time, Nvidia has partnered with Unitree Robotics to launch a new-generation humanoid robot reference design, H2+, also known as the Isaac GR00T system. In October, Nvidia announced an expansion of the Cosmos world model platform, adding training tools, simulation checkpoints and Cosmos Data Factory blueprints to help physical AI developers accelerate data pipeline construction.

At this year's GTC conference in Berlin, Nvidia CEO Jensen Huang explicitly said that "the next wave is physical AI," which also triggered a wave of discussion about physical AI in China at the time.

Third, the foundation models laid out by big tech are beginning to move toward the physical world.

In September, OpenAI released GPT-6 Astra. The focus was not on building a dedicated embodied model from scratch, but on bringing the accumulated capabilities of general models in perception, reasoning and planning into the robotics field, which triggered industry discussion about whether general models constitute a dimensionality-reduction attack on specialized models.

For a time, the core moat and room to survive of dedicated embodied intelligence models came under scrutiny. The core of this shock is not that Astra can immediately replace specialized models, but that it shook the logic that had previously supported the valuations of independent embodied "brain" companies. Some overseas commentary said: "The OpenAI of the robotics field will be OpenAI itself, not Physical Intelligence or Skild."

Other big tech companies also have related layouts. Alibaba released three major models—the VLA manipulation model Qwen-RobotManip, the VLN navigation model Qwen-RobotNav, and the world model Qwen-RobotWorld—attempting to use one model system to cover the manipulation, navigation and physical law prediction of robots of different forms. In addition, Anthropic released the "Model Hardware Standard" (MHS), which overseas media interpreted as its first public entry into the fields of embodied intelligence and physical AI, with its entry point being to set interface standards for AI-controlled physical devices.

▌The A-share physical AI industry chain: most are still "shovel sellers"

In the secondary market, in the just-passed October, the A-share physical AI industry chain set off multiple rounds of limit-up waves. At present, A-share layouts in physical AI basically unfold along the chain of "perception—physical simulation/world models—computing power—embodied bodies and supporting facilities—industrial AI/industry scenario applications." Most companies are still in the position of "selling shovels." Very few A-share companies truly do physical AI simulation platforms and world models; Suochen Technology is one of them.

In March this year, Suochen Technology released the world physical model "Yingzao·Wanxiang" and the agent "Dasheng," benchmarking Nvidia's Omniverse/Isaac Sim system. In the first half of 2026, the "Kaiwu" physical AI business had achieved revenue of 6.59 million yuan, accounting for 9.06%; R&D investment was 84.8 million yuan, up 67.53% year on year, accounting for 116.59% of operating revenue. It is worth noting that the company's total revenue in the first half of the year was 72.74 million yuan, with a net loss attributable to shareholders of 78.66 million yuan. The company also clearly stated in its announcement that the physical AI business is currently "in the early stage of layout, and commercialization is uncertain."

Xie Jun, investment manager at Hang Seng Qianhai, believes that physical AI extends AI capabilities into the real physical world and, relying on world models, simulation, perception and execution hardware, opens up new demand for computing power and continues the computing power narrative. He believes that physical AI is currently in the early growth stage from technical validation to small-batch pilot delivery, and he is optimistic about three major directions: simulation and world model toolchains, perception hardware, and execution components.

For physical AI to land, "foundations" such as perception, computing power, components and simulation tools are indispensable, which is also the consensus of many practitioners in interviews. But for the secondary market, in the conceptual carnival it is even more necessary to distinguish who is "truly laying out" and who is "riding the hot topic." What must be guarded against is that before embodied intelligence complete machines usher in large-scale volume ramp-up, it is difficult for the upstream of the industry chain to have substantive performance support. From concept to orders, the industrial long run of physical AI has only just begun