Embodied AI talent gap reaches one million, interns offered monthly salaries of 100,000 yuan

[[**Embodied AI financing and hiring demand explode, with a talent gap of one million; companies are poaching talent with high salaries, interns can earn up to 100,000 yuan a month, and the annual salary of chief scientists is rumored to reach the hundred-million-yuan level; but high financing pushes up costs, and the risk of industry divergence rises.**

**Key points**

**1. Job demand surges while talent supply is severely insufficient**

From January to July 2026, newly posted positions increased 6 to 7 times year-on-year, with some humanoid robot companies up 831.88%; the talent gap is about 1 million, and the supply-demand ratio for R&D positions is 1:10.

**2. High-salary poaching extends from fresh PhDs to interns and chief scientists**

Fresh PhDs earn about 1.5 million yuan a year, with outstanding ones at 3 million to 5 million yuan; some startups offer interns a monthly salary of 100,000 yuan; the annual salary of chief scientists is rumored to reach up to 124 million yuan, with another rumor of 200 million yuan for recruitment.

**3. Capital FOMO pushes up financing and employment costs**

In 2026, core tracks raised about 935 rounds and 135.2 billion yuan, tripling in three years; even very early-stage companies can obtain high financing and high valuations, and spending without regard to cost may become a burden.

**4. Long cycles and heavy accumulation, with the risk of industry divergence emerging**

Embodied AI requires continuous refinement in areas such as joint modules, motion control, and real-machine generalization; if companies focus on financing stories and neglect technological breakthroughs, they are most likely to be eliminated when divergence comes. 2026-10-25 15:45 Interns can now earn 100,000 yuan a month too, ChinaVenture quick look

"Recently I've been busy poaching people, and even interns are being offered a monthly salary of 100,000 yuan."

A headhunter friend said this recently. If this had been said two years ago, no one would have believed it; but today, probably no one would dare not believe it.

Since the beginning of this year, embodied AI has undoubtedly been the brightest star. Both financing events and financing scale have made it the undisputed "money magnet." However, the money is in place, but the people are not. Recently I heard a saying: because the financing amount is too high, some embodied AI companies don't have enough manpower. This is no joke, but the real situation an investor confided to me.

The most extreme example is a leading humanoid robot company that offered a price of 124 million yuan to poach a "chief scientist." And in the second half of the year, this "poaching" trend has become increasingly fierce. In multiple group chats, recruitment information for "technical positions" posted by embodied AI companies can be seen.

Embodied AI is short of manpower

The most direct reflection of embodied AI companies not having enough manpower is the "number" of job postings.

According to the "2026 Autumn Recruitment Workplace Insight Report" by the Maimai Gaopin Talent Think Tank, from January to July 2026, the number of newly posted positions in embodied AI increased 6 to 7 times year-on-year. Among them, the recruitment volume in the humanoid robot sub-track grew prominently, with some leading humanoid robot companies' newly posted positions increasing by as much as 831.88% year-on-year, involving multiple core positions such as algorithm engineers, mechanical structure engineers, and motion control engineers.

Despite the high growth in demand, supply has not kept up. Survey data from Beijing Institute of Technology shows that the current talent gap in embodied AI across industries and organizations is as high as 1 million. Other data shows that the supply-demand ratio for high-performance computing engineers plummeted from 0.99 in 2025 to 0.26, equivalent to changing from "1 person per position" to "4 positions competing for 1 person"; the supply-demand ratio for embodied AI R&D positions is as high as 1:10, and the "talent shortage" phenomenon is very prominent.

Against this background, the cost of poaching has also risen. A vice president of a leading humanoid robot company once said that leading general-purpose talent is now very scarce. "The cost for a company to hire a fresh graduate in embodied AI is also quite high: a PhD requires at least about 1.5 million yuan in annual salary, and outstanding ones require 3 million to 5 million yuan in annual salary."

This means that the already unbalanced talent supply-demand contradiction is further intensified. A set of data shows that starting in 2025, the Ministry of Education only newly added the "embodied AI" undergraduate major, and across the country, only 9 institutions have been approved, so the number of corresponding talents delivered each year can be imagined.

At the same time, the cycle of talent cultivation has once again accelerated the expansion of this gap. An AI industry investor admitted that training a qualified mechanical, automation, or precision electromechanical engineer takes several years to gain experience in offline hands-on work, production line debugging, and project refinement.

"For companies in urgent need of technical R&D positions, fresh graduates simply cannot get started quickly in the short term like general algorithms."

In addition, embodied AI is also an emerging interdisciplinary field. There are also many people in the market who have transitioned from fields such as autonomous driving, new energy vehicles, and traditional intelligent manufacturing, but how many will truly stay is unknown. Moreover, embodied AI tests the ability to build systems from 0 to 1 and the ability to land on real machines, rather than a single technology stack, which to a certain extent also raises the threshold for leading talent.

Interns can also earn 100,000 yuan a month

In fact, this wave of talent shortage already appeared in the spring of this year. It's just that by autumn, the trend has become more obvious.

Recently, a person related to the embodied AI field clearly told me, "Recently, she has been traveling on business for several weeks in a row, so busy that her feet hardly touch the ground." Another startup embodied AI company that has completed several billion-yuan financing rounds in succession also told me, "The company is hiring in large numbers recently. If you find suitable people, please recommend them."

Headhunter Icy confirmed this to me. In her view, there have recently been many demands like this of "thirsting for talent." "Just for motion control experts, there are several companies looking, which shows the scarcity of industry talent."

In Icy's view, "now it's not just the leading embodied AI companies looking for 'experts'—embodied AI startups are also going all out." Every "chip" one could think of—lightning-fast interviews, hefty signing bonuses, executives personally lobbying, promises of customized projects—has been put on the table.

It is understood that, to lock in potential talent, some startups have introduced a policy of "paying full-time salaries during internships." That is, as long as candidates agree to join after graduation and start interning half a year or even a year early, they can receive a full-time salary every month. Some have even gone further, directly offering interns a high salary of 100,000 yuan per month.

What does a 100,000-yuan monthly salary mean? Within the embodied AI circle, it is already close to the monthly salary level of a seasoned algorithm engineer. Two years ago, for the same position, an intern's daily wage might still have been in the range of a few hundred yuan, and those who could get a monthly salary of ten thousand yuan were rare. Today, a doctoral student who has not yet graduated, as long as their research direction hits the right spot, can have internship income comparable to a full-time salary...

More critically, this is not an isolated case. A headhunter told me that similar offers are no longer a secret among leading embodied AI companies: "If you don't offer this price, the person may go next door." For this reason, outside discussion of these numbers has gradually shifted from the initial "is it real?" to "how high can it go?"

And that "high" is already written into the price tag for "chief scientist." Icy said that the current salary ceiling in embodied AI is mainly concentrated in three positions: "embodied AI chief scientist," "senior large-model algorithm expert," and "world-model algorithm expert." In the first half of this year, just to compete for "chief scientists," the above-mentioned leading humanoid robot company offered annual salaries starting at 15 million yuan and up to 124 million yuan, directly benchmarked against top scientists at OpenAI and Meta.

Recently, screenshots from an internal group chat leaked by a newly emerging leading embodied AI company also showed that a controlling shareholder called for recruiting a "chief scientist" with an annual salary of "200 million yuan."

Faced with all this, Icy said the industry is changing too fast. "A year ago, almost no one was looking for the position of 'world-model algorithm expert,' but now it has become a hot job. As for what it will be like next year, no one can say for sure, but in the short term, this wave of talent poaching indeed shows no sign of cooling down."

Amid the talent scramble, internal strength still needs to be honed

Looking at the outbreak of this "talent scramble," it is the result of capital's FOMO sentiment and a rush of money flooding in. As of now, in 2026, the core embodied AI track has seen about 935 financing deals, corresponding to 135.2 billion yuan, which is equivalent to tripling in three years compared with the previous two years.

Against this backdrop, an investor told me that some very early-stage companies, relying on a founding team that looks decent, can also obtain relatively high financing and good valuations.

But such valuations and financing ultimately have to be repaid. Previously, a leading investor worried that, in all likelihood next year, trillions or even tens of trillions of funds will require investment institutions to deliver returns and DPI to LPs. What will happen then, and how companies and investment institutions should respond, are all issues that need to be considered.

What is worth being vigilant about is that when a company completes multiple financing rounds in succession within half a year, the mindset of the founding team can also easily change. "It's like, after suddenly having several hundred million on the books, hiring can become cost-insensitive, office space can be upgraded immediately, and even marketing activities start being carried out on a grand scale. But these actions are not equivalent to technological breakthroughs or orders landing."

In this investor's view, sometimes money coming too easily is not a good thing; instead, it makes it easy to neglect the most important fundamentals. "Embodied AI is a long-cycle track that emphasizes accumulation. From joint modules to motion control algorithms, and then to the generalization of real machines in real scenarios, every item requires careful polishing and hard work on internal strength."

"If the team puts too much energy into financing rhythm, valuation management, and how to tell stories externally, and does not truly settle down to tackle technical challenges, then once the industry diverges, it may be the first to be thrown off the bus."

A more realistic consideration is that some companies may push up fixed costs because of excessively high financing amounts, but whether these people are truly effective is unknown. Once the subsequent financing pace slows, or technological progress falls short of expectations, these costs will quickly become a burden.

From the beginning of the year to the end, embodied AI has remained hot until now, and this is almost the hottest time for talent compensation in this industry to climb. But the rule of history is that behind the prosperity of any industry, it almost always starts with "universal" optimism and ends with a "divergent" outcome—and this divergence is not only among these companies, but also among the people lifted high by this wave.

Those doctoral students holding internship offers with a 100,000-yuan monthly salary, those founders offering 200 million yuan annual salaries to grab chief scientists, and those practitioners who travel continuously and hardly touch the ground are actually all being lifted by the same wave and are all standing on this ship. No one knows which way the tide will recede, but what is certain is that when divergence comes, not everyone will be able to stay on deck. AI industry signal channel: Frontier Technology. This content is authorized by the author for publication, and the views represent only the author and do not represent Huxiu's position. If you have objections or complaints about this manuscript, please contact tougao@huxiu.com. Everyone who is changing and wants to change the world is in the Huxiu APP[]