2026-06-14
China's humanoid robot sector moves toward public markets: EngineAI, Unitree, and Linkerbot advance IPO filings as embodied AI scales
— views
By June 14, 2026, EngineAI, Unitree, and Linkerbot were advancing IPO filings, marking China's humanoid-robot sector's shift from prototypes to public companies. EngineAI reportedly produces one humanoid every 15 minutes — industrial-scale output ahead of listing.
What is happening
By the week of June 14, 2026, at least three Chinese humanoid robot companies were actively advancing public listing filings: EngineAI, Unitree Robotics, and Linkerbot. The timing marks a sector-level inflection point: China’s humanoid-robot space, which was largely at the demo-and-prototype stage 18 months ago, has reached the production throughput and revenue visibility needed for public-market investor scrutiny.
The most striking data point is EngineAI’s reported production rate: one humanoid robot every 15 minutes. If accurate at scale, that is roughly 100 units per day per production shift — an output level that crosses from “we are making robots” into “we are manufacturing robots.” For context, most U.S. humanoid companies are still delivering units in hundreds per year to enterprise beta customers.
The three companies
EngineAI is Shenzhen-based and makes the PM01 humanoid, a 62kg bipedal robot targeting logistics and light manufacturing. Its production-line claim of one unit per 15 minutes would put it at meaningful industrial throughput. EngineAI has focused heavily on cost reduction through domestic component supply chains — key joints, actuators, and sensors sourced within China, avoiding the U.S. supply-chain exposure that complicates scaling for some competitors.
Unitree Robotics is probably the most globally known Chinese humanoid maker, best recognized for its viral G1 and H1 demos and its Go series of quadruped robots. Unitree has been selling quadrupeds commercially for years and has the closest-to-established revenue track record of the three. Its humanoid products (G1, H1) are available for purchase at prices significantly below Western competitors: the G1 starts at approximately $16,000, which is substantially cheaper than comparable U.S. or European humanoids. That price point, if maintainable at volume, is a structural threat to higher-cost competitors.
Linkerbot is less publicly known outside China but has been positioned in the light-manipulation and service-robot category, targeting applications in healthcare, hospitality, and retail in addition to light industrial tasks.
Why Chinese humanoid IPOs matter globally
It establishes a valuation anchor. U.S. humanoid companies are currently valued in venture markets with limited public comparables. Chinese public listings will create the first substantial public-market pricing for humanoid robotics, which will cascade back into how U.S. and European VCs mark their private portfolio companies.
It accelerates the cost-down race. Chinese manufacturers at public scale will face relentless pressure to reduce unit cost and increase margin. That competitive pressure benefits end customers everywhere by pulling the price floor of humanoid robots down faster than it would fall in a market with only private players.
It forces a technology reckoning. Public companies must disclose production numbers, defect rates, and customer contracts in a way that venture-backed companies do not. The actual capability gap between Chinese and Western humanoids will become more legible once the filings are public.
Production throughput as a competitive moat
The “one unit per 15 minutes” EngineAI claim deserves scrutiny. There are several interpretive scenarios:
- True sustained throughput — the production line is running at that cadence continuously. This would put EngineAI at approximately 96+ units per day per shift, or potentially thousands per month at multi-shift production.
- Peak-rate claim during a PR run — a single shift operating at maximum throughput for a demonstration or press visit. Real capacity utilization is lower.
- Partially assembled units on a line — final assembly at that speed, with subcomponent manufacturing times not counted.
Regardless of which interpretation is correct, the number is directionally meaningful: EngineAI has a functioning production line operating at industrial cadence. That is different from a “we can make them” claim.
The price gap problem for U.S. and European competitors
Unitree’s G1 at ~$16,000 is an inflection-point number for the industry. For reference:
| Product | Company | Approximate price |
|---|---|---|
| G1 | Unitree (China) | ~$16,000 |
| H1 | Unitree (China) | ~$90,000 |
| 4NE-1 | Neura Robotics (Germany) | Not public |
| Figure 02 | Figure AI (US) | Not public (estimated $100K+ enterprise) |
| Atlas | Boston Dynamics (US) | Not publicly sold |
A humanoid at $16,000 that can do light material handling begins to compete economically with human labor in a range of warehouse and manufacturing scenarios, especially in high-wage markets. The question is whether Unitree’s current production quality and software reliability are high enough to deploy unsupervised at that price point — reports from enterprise customers are mixed.
Practitioner note
- IPO filings are not IPOs. Filing is a step in a process that can take 12–24 months and can be withdrawn. Watch for actual listing dates and the disclosed revenue and unit numbers in filings, not the filing announcement itself.
- The regulatory environment matters. Chinese humanoid companies doing public listings in Hong Kong or domestic A-share markets will face different foreign-investor access constraints than if they listed in the U.S. (which CFIUS would complicate). The venue choice will determine how much Western capital can flow in.
- Supply chain localization is a double-edged sword. Sourcing key actuators domestically reduces cost and avoids export-control risk — but Chinese actuator technology, while improving rapidly, is still generally considered behind Japanese (e.g., Harmonic Drive) and Korean precision suppliers. That gap may close, but it is a current quality risk.
Under-considered angle
The narrative around Chinese humanoid robots defaulting to “geopolitical threat” misses the more immediate structural question: China is about to establish public-market proof of concept for humanoid robots as a mass-manufactured product category. If EngineAI’s production numbers hold up and the IPO succeeds, the assumption that humanoids are 5–7 years from manufacturing scale will be retested. U.S. and European policymakers who assumed time to respond may find the window is shorter than expected — not because of espionage or IP theft, but because China has run the manufacturing learning curve faster than the rest of the industry anticipated.