Cutting-Edge Technologies
NextFin News - China's humanoid-robot rush has reached the IPO gate, and the gatekeeper is tightening the lock. Noetix Robotics, the Beijing startup whose machines danced on national television and whose Series B was led by battery giant CATL's investment arm, has confidentially filed for a Hong Kong listing that could raise several hundred million dollars and debut as soon as 2027. The filing arrives just as China's securities regulator raises the bar for embodied-AI listings — turning Noetix's debut into the sector's first real stress test.
The central question is not whether Noetix can sell a filing. It is whether China's most hyped hardware sector can survive contact with public-market discipline after a funding boom that valued demos ahead of earnings. The answer will set the template for at least two dozen other humanoid companies waiting in the Hong Kong queue.
The Filing, the Backer, and the Nine-Round War Chest
Noetix Robotics Beijing Technology Co. Ltd. — formally known as Songyan Power Beijing Technology Co. — submitted a confidential application to list in Hong Kong, according to people who asked not to be identified discussing private information. The company is targeting a public debut in 2027 with a raise in the hundreds of millions of dollars. Confidential filings, which Hong Kong began permitting for technology companies in May 2025, let pre-profit startups probe investor appetite without immediately exposing their books.
The headline name on the cap table is Contemporary Amperex Technology, better known as CATL. Its affiliated industrial investment platform, Chendao Capital, led Noetix's Series B round of nearly 1 billion yuan (about $145.9 million), announced on March 2, 2026. That was the company's ninth financing round since Jiang Zheyuan founded it in September 2023, having left a Tsinghua University doctoral program to commercialize legged-robot motion control.
Nine rounds in under three years is the story in miniature. Noetix has built its reputation on price disruption rather than technical maximalism. Its Bumi model, standing 94 centimeters tall, is marketed as the world's first high-performance humanoid under 10,000 yuan — roughly the cost of a midrange smartphone. The N2, an 18-degree-of-freedom athlete robot, can run at up to 3.2 meters per second and perform flips and sprints. The larger E1, at 1.36 meters with 21 degrees of freedom, targets embodied-intelligence interaction with synchronized speech and facial expression. The company's robots then appeared on CCTV's Spring Festival Gala on February 16, 2026, reaching hundreds of millions of viewers and converting a hardware startup into a household name almost overnight.
"Our success at this year's Games reflects the progress we have made and the path we are committed to," said Zheyuan Jiang, chairman of Noetix Robotics, after the company's robots won two gold medals and a silver at the Global Humanoid Robotics Games. "Humanoid robots are more than an engineering breakthrough. They are a bridge between technology and human life. At Noetix, we are focused on moving robots beyond the laboratory and into everyday settings, where they can become trusted companions and assistants."
The quote captures the sector's ambition — and the gap between that ambition and a public-market reality that demands audited revenue, not podium finishes.
A Queue of Two Dozen Filers Meets a Regulator With a Filter
Noetix is not alone at the gate. At least two dozen humanoid-related embodied-AI companies have filed to list in Hong Kong, according to people familiar with the regulator's thinking. China now has well over 100 humanoid companies, all riding a national policy push that has named embodied intelligence a priority future industry in two consecutive government work reports. Investment in the sector hit 47.09 billion yuan ($6.95 billion) in the second quarter of 2026 alone — more than double the first quarter and over six times the same period a year earlier, according to industry data provider Xiniu.
But the door is narrowing. China's securities regulator wants local embodied-AI startups seeking to go public to meet three specific criteria, and even under the more lenient reading — satisfying two of the three — it remains unclear how many of the filers can qualify. The criteria themselves have not been publicly disclosed, and the China Securities Regulatory Commission did not respond to requests for comment. The Hong Kong stock exchange declined to comment. Mainland Chinese companies seeking a Hong Kong listing still require the CSRC's approval, which gives Beijing an effective veto over the entire pipeline.
This is the mechanism through which the sector's repricing is being enforced. Beijing is not trying to kill embodied AI — it has made the technology a national priority. It is trying to separate companies that can manufacture and sell at scale from those that can only demonstrate. A confidential filing lets a startup start the clock; the CSRC's blessing determines whether the clock ever reaches a listing date. The intervention converts a funding-cycle problem into a structural filter.
The Public-Market Warning Shot Already Fired
Investors already have their first live experiment, and the read is sobering. Unitree Robotics, the industry's poster child, won a regulatory fast-track to its Shanghai STAR Market listing on August 19, 2026, raising about 6.1 billion yuan ($905 million) as shares closed more than 460 percent higher on debut at 845 yuan. The stock briefly touched 1,100 yuan, pushing the company's market value toward 445 billion yuan. A day later, Unitree's founder Wang Xingxing cautioned that commercialization beyond dancing robots remained years away.
The hangover was swift. Unitree's shares have since fallen roughly 39 percent from their debut close to about 528 yuan, erasing tens of billions of yuan in market value from the peak. The company did something almost no Chinese humanoid rival has done — it is profitable, with 2025 operating income of 1.708 billion yuan, up 335 percent year over year, and net profit up 674 percent. If the sector's strongest fundamental story cannot hold its IPO pop, the bar for the rest of the field is that much higher.
The warning light is brighter still in Hong Kong itself. UBTech Robotics, China's largest humanoid maker and the sector's only major Hong Kong-listed pure play, has lost more than 40 percent of its value this year. As of October 22, UBTech shares traded at 65.10 Hong Kong dollars, down 3.34 percent on the day, with a market capitalization of about 31.2 billion Hong Kong dollars — less than half of the company's 52-week high of 156.40 Hong Kong dollars. The fundamentals tell the same story as the stock: second-quarter 2026 revenue of HK$732.9 million rose 115.3 percent year over year, yet the company still posted a net loss of HK$179.9 million, a net margin of negative 24.5 percent.
Sales are growing fast. Profits are not following. That is the entire sector in one margin line.
Why This Is a Structural Filter, Not a Cyclical Cool-Down
The easy read of Noetix's filing is that it is a cyclical cash grab: valuations peaked, private capital is turning cautious, and founders are rushing to monetize before the window closes. That reading is partly right but stops at direct causality. The deeper story is structural, and it is about a business model being repriced.
China's humanoid sector was built on a funding model that rewarded speed over unit economics. Startups raised billion-yuan rounds on the promise of a consumer market that does not yet exist at scale. A robot priced at $1,370 to $5,500 can win headlines, gala slots, and competition medals, but it cannot yet pay for the R&D, the service network, and the safety engineering that real deployment requires. The model — cheap hardware, deferred profitability, endless capital — is what is being repriced, not merely the sentiment around it.
Three pieces of evidence point to a regime shift rather than a cycle that will revert on its own. First, the regulatory criteria, once established for a sector, do not get walked back when sentiment improves; they become part of the market's architecture. Second, the funding surge is itself the bubble signal: second-quarter investment up six times year over year is the kind of parabolic capital inflow that precedes discipline, not the kind that sustains it. Third, the public-market comparables are already speaking — Unitree down 39 percent from its debut close despite profits, UBTech down more than 40 percent this year despite 115 percent revenue growth. When the best stories in a sector cannot hold their listings, the problem is not temporary positioning.
The implication for Noetix is double-edged. With nine financing rounds behind it, a CATL backer that understands industrial scale, and products already in the market, the company is better positioned than most of the two-dozen filers. But even a well-connected applicant must now prove what no Chinese humanoid company has yet proven in public markets: that it can make money, not just robots. The CSRC's criteria — whatever they turn out to be — are the test.
The Second-Order Effect: What the IPO Means for CATL
The angle most investors will miss is what a Noetix listing does for its marquee backer. CATL's shares have fallen sharply in 2026 — trading at 221.82 yuan on October 22, down 2.67 percent on the day, with a market capitalization of about 978 billion yuan, roughly 53 percent below its 52-week high of 468.75 yuan. The electric-vehicle battery market is maturing, growth is slowing, and the company is searching for its next curve.
Humanoid robots are power-hungry machines, and CATL has already moved beyond passive investment into active deployment. The company launched what it called the world's first humanoid-robot-powered battery-pack production line at its Luoyang base, using its Xiaomo humanoid for tasks such as battery connector insertion. It has also collaborated on heavy-duty humanoid robots for factory use. A successful Noetix listing would give CATL a marked-to-market valuation for its robotics exposure and a strategic foothold in a downstream consumer of battery technology.
But the reverse is also true, and it is the second-order risk. If Noetix stumbles at the listing gate — rejected, delayed, or priced at a steep discount — it becomes another mark against CATL's diversification bets at a time when the core battery business is already under pressure. The IPO is a two-way mirror: it tests Noetix's viability, and it tests whether CATL's robotics thesis is an investment or an option. For a battery maker trading at a fraction of its peak, the option value matters.
The Counter-Thesis, and the Signal That Would Break This One
The strongest case against this skeptical read is that regulators are not blocking listings — they are curating them. Beijing wants national champions, not a graveyard. Unitree received a record 104-day fast-track approval from filing to listing; the criteria, even at two of three, are meant to ensure quality rather than to shut the door. Under this view, Noetix is not a bubble pop waiting to happen but a well-connected contender that has already cleared the hardest hurdle: attracting serious capital and a serious industrial backer.
That argument has force, but it rests on a hidden premise — that Noetix actually meets the criteria. The criteria have not been published, and sources say it is unclear which, if any, of the two-dozen filers can satisfy them. A fast-track for one company is not an open road for twenty-four. The counter-thesis works only if Noetix's revenue and shipment base is materially stronger than the sector average, and that is precisely what a confidential filing keeps hidden from investors.
The falsifying signal is concrete. If Noetix receives CSRC approval and prices its Hong Kong IPO at or above its last private valuation, with a strong first-day pop and a stable aftermarket through the first quarter after listing, then the bubble thesis is wrong and the sector has found its public-market footing. If the filing lapses, is rejected, or prices at a steep discount and then trades down like Unitree, the repricing has further to go. Watch the prospectus revenue line and the CSRC decision date — those two data points will settle the argument.
What to Watch Next
Three signals will tell the story over the next six to twelve months. First, the CSRC's decision on Noetix's overseas listing application — approval, rejection, or silence. That single data point matters more than any gala appearance. Second, the revenue and shipment figures Noetix discloses when its prospectus becomes public; anything short of meaningful commercial scale will confirm the sector's fundamentals gap. Third, the performance of Unitree and UBTech through the end of 2026; if both remain underwater, investor appetite for a third humanoid listing will be thin regardless of the demo reels.
Short term, the filing is a sentiment event in a sector that has already been punished. Medium term, it is a fundamentals test that most of the field is not built to pass. Long term, the survivors of this regulatory filter — if any emerge — will define what Chinese humanoid robotics actually becomes once the hype cycle ends.
Noetix's IPO is not the beginning of China's humanoid-robot public-market era. It is the first stress test of whether that era can exist at all — and the regulator, not the market, is holding the pass-fail button.
Explore more exclusive insights at nextfin.ai.
