AI News

Primary market reluctant to openly discuss physical AI; embodied intelligence financing falls back to 33.335 billion yuan in Q3

HuxiuUpdated Nov 15 09:28

AI summary

Multiple investment institutions are no longer willing to publicly discuss physical AI; after one institution shared publicly, it received inquiries from LPs about investment risk. From July 2025 to June 2026, the domestic embodied intelligence primary market saw a cumulative 503 financing deals totaling over 96 billion yuan; in Q3 2026 it was about 33.335 billion yuan, a pullback from the first half of the year when each quarter exceeded 46 billion yuan. Unitree's market value was about 444.9 billion yuan at the open on its first trading day, fell to about 163.5 billion yuan in mid-October, and rebounded to about 281.5 billion yuan on October 28; the market value swings prompted LPs to reassess risk, and the focus of investors' questions also shifted from technology demonstrations to real-world scenario deployment.

Why read this

AI score 56

From three layers—financing data, Unitree's market value fluctuations, and institutional statements—this presents the primary market's shift toward physical AI from scrambling for share to reassessing risk.

Topics

Tags

Body · AI translation

[[**Primary-market financing for embodied intelligence remains hot, but institutions are collectively keeping quiet; LPs are reassessing risk due to sharp fluctuations in Unitree's market value; financing is slowing, investors are focusing more on real-world deployment, and they are worried about large-model companies entering the field.**

**Key points**

**1. Investment institutions avoid publicly discussing physical AI**

Many institutions that had previously invested in or engaged with projects are unwilling to discuss physical AI publicly, and some internally no longer want much discussion of it; after one institution shared publicly, it received inquiries from LPs about risk.

**2. Financing scale remains large but fell in Q3**

From July 2025 to June 2026, there were 503 deals cumulatively, exceeding 96 billion yuan; in Q3 2026, about 33.335 billion yuan, down from more than 46 billion yuan per quarter on average in the first half.

**3. Sharp fluctuations in Unitree's market value prompt LPs to reassess risk**

Unitree's market value was about 444.9 billion yuan at the opening on its first day, fell to about 163.5 billion yuan in mid-October, and rebounded to about 281.5 billion yuan on October 28; the rise and fall in market value has made LPs more cautious.

**4. Investment focus shifts from grabbing share to real-world deployment**

Previously, institutions scrambled for share, and valuations crossed thresholds within a week; now technical demonstrations are not enough, and real scenarios and business progress have become the focus of questioning.

**5. The entry of large-model companies may disrupt the valuation logic of startups**

This article is from the WeChat public account: Cailianshe, author: Yu Shiqi

These institutions are not bystanders to embodied intelligence. Some of them have long participated in related investments, while others continue to engage with financing projects in this field. Some institutions still recognize the long-term development space of physical AI, but they already have more concerns about the current investment heat and public expression.

LPs begin to re-examine physical AI

According to incomplete statistics by a reporter from the Science and Technology Innovation Board Daily, from July 2025 to June 2026, 503 financing deals occurred cumulatively in the domestic primary market for embodied intelligence, with a total amount exceeding 96 billion yuan, excluding IPOs and M&A. On average, more than one financing event occurred every day.

With a large amount of capital flowing in, the financing scale and valuations of leading companies have continued to climb. For investment institutions, whether they can enter this direction in time and obtain shares in high-quality projects has become the focus of competition. In this field, which is believed to have the potential to produce a new generation of tech giants, missing out on leading projects means possibly losing investment opportunities for years to come.

English text is being prepared.

But now, many investment institutions have begun to reconsider how to talk about this direction publicly.

A person from an investment institution told a reporter from the Science and Technology Innovation Board Daily that the institution had previously participated in physical AI-related activities and publicly shared technology and industry prospects, and afterward received inquiries from LPs. What capital providers care about is no longer just how big the future of the technology is, but also the investment risk of institutions in this direction.

The person still recognizes the long-term development space of physical AI. In his view, physical AI has certain technological connections with fields such as robotics and autonomous driving, and its development prospects need to be observed over a longer time cycle. But for investment institutions, being long-term optimistic about the technology is one thing, and how to invest in the current market environment and how to express themselves externally is another.

Other investment institutions also have their own concerns. Some institutions had already invested in embodied intelligence but are now unwilling to discuss this direction publicly; others still recognize the long-term value of related technologies but have reservations about the current investment heat and corporate valuations.

Unlike the aforementioned institutions that directly mentioned LP inquiries, when the Science and Technology Innovation Board Daily reporter communicated with multiple investment institutions, he found that Unitree Technology's market value performance after entering the public market was a topic repeatedly mentioned.

Previously, the valuations of embodied intelligence companies kept rising through rounds of financing. Especially for popular projects pursued by capital, the highest quote often became an important reference for the valuation of a new financing round. The secondary market is different; public trading keeps prices continuously tested by the market.

On August 19, its first day of listing, its opening market value was about 444.9 billion yuan, then in mid-October it once fell to about 163.5 billion yuan, and by October 28 it rebounded to about 281.5 billion yuan. In just over two months, its market value experienced sharp rises and falls.

It is worth noting that it is not only investment institutions paying attention to this change, but also the LPs behind them. The price fluctuations of leading companies in the secondary market have led some LPs to begin re-examining embodied intelligence as an investment direction. For them, a track once held in high expectations now also requires a more cautious assessment of investment risk.

And the experience of the aforementioned investment institution receiving LP inquiries also gives this change a more concrete manifestation: capital providers' attention to this field is no longer only about investment returns, but has also extended to how GPs express their investment judgments externally.

From grabbing share to recalculating prices

How hot was embodied intelligence investment last year? A person from an institution that had participated in early investment in a leading embodied intelligence company told the Science and Technology Innovation Board Daily reporter a story.

At the time, this institution had clear limits on project valuation and needed to complete the investment within 1 billion yuan. But with leading projects pursued by multiple institutions, whether they could catch the financing window almost determined whether the institution could obtain an investment share.

This institution ultimately made it in time. But just one week later, the project valuation crossed its investment threshold. At the new price, this institution could no longer enter. For investors at the time, one week earlier or one week later could mean completely different investment results.

Such stories are one side of the previous wave of embodied intelligence investment. The financing windows for leading projects keep shortening, and valuations are rising rapidly; investment institutions must not only judge the direction of technology but also compete for limited investment shares.

In the first half of this year, embodied intelligence remained one of the most closely watched investment directions in the primary market. Entering the second half of the year, financing amounts began to decline somewhat.

According to incomplete statistics by the Science and Technology Innovation Board Daily reporter, in the third quarter of 2026, at least 185 financing deals occurred in the domestic embodied intelligence sector, with financing amounts of approximately 33.335 billion yuan. Compared with the financing scale of more than 46 billion yuan per quarter on average in the first half of the year, the financing amount in the third quarter decreased.

A robotics company that the Science and Technology Innovation Board Daily reporter had previously contacted completed three consecutive financing rounds this year and has relatively ample capital reserves. This company has a differentiated technology route and special application scenarios, and is not a typical general-purpose humanoid robot project.

The head of the company told the Science and Technology Innovation Board Daily reporter that early market expectations for embodied intelligence were overheated to a certain extent. By the most recent financing round, although funds still arrived relatively quickly, investors' attitudes had become more cautious than before.

An FA professional still in contact with embodied intelligence projects also described a similar feeling.

From the internal advancement of projects at institutions to feedback from external investors, this FA professional felt that the financing pace of embodied intelligence projects had slowed somewhat. In their observation, the hottest financing window had passed; teams within institutions were still advancing related projects, but investors were no longer as eager to act as before.

The pace of investment institutions' moves has changed somewhat. At the same time, the Science and Technology Innovation Board Daily reporter noticed that some leading companies that had previously obtained large financing amounts have recently begun to differ in what they emphasize externally.

Galaxy General, which completed 2.5 billion yuan in financing in March this year, is an example. Compared with previous introductions centered on embodied large models, technical achievements, and research papers, entering the second half of the year, this company began to show more progress in the application of robots in real scenarios such as smart pharmacies.

In the Science and Technology Innovation Board Daily reporter's recent exchanges with primary market institutions, changes in investors' focus could also be felt. Model capabilities and technological breakthroughs remain important, but for embodied intelligence projects, technical demonstrations alone are no longer enough. How many real scenarios robots have actually entered and what stage actual business advancement has reached have become questions investors further probe.

Recently, news that ByteDance's Seed team is recruiting robot-related talent and that DeepSeek founder Liang Wenfeng is said to be learning about embodied intelligence companies has also attracted attention from some investors. As large model capabilities continue to break through, it remains difficult to judge how much these large AI companies with computing power, capital, and technical resources will change the competitive landscape of embodied intelligence in the future.

A founder in a robotics niche track also mentioned similar concerns in a recent exchange. In his view, large model companies themselves have more ample R&D resources. If they further enter physical AI, especially the field of embodied intelligence brains, this may not be a favorable change for startups currently focused on this direction as their core business.

When large AI companies begin to pay attention to embodied intelligence brains, what investment institutions need to face is no longer just whether startups' current technical capabilities match their valuations. How long those technologies once considered to have a leading advantage can maintain it in the future also affects investors' judgment of project value. AI industry signal channel: Frontier Technology This content comes from the internet, original link; the views represent only the author and do not represent Huxiu's position. For copyright issues, please contact hezuo@huxiu.com, and we will verify and handle them promptly. Everyone who is changing and wants to change the world is in the Huxiu APP[ ]

Source

Huxiu