At three in the morning, a sorting centre outside Beijing runs at full speed. Parcels of every shape and material come down the line; a robot arm flips each one so the address label faces up, reads it, and drops it into the right chute. No one is watching to see whether it works. It has to work, because the vans leave at five.
The company behind those arms, Robot Era (星动纪元), has become one of the more closely watched private companies in Chinese robotics — not because of a stage demo, but because the work is boring, repetitive and paid for. Bloomberg has reported that the company is working with advisers on a Hong Kong listing that could raise between US$800 million and US$1 billion. Robot Era has not confirmed terms, and the discussions could still change.
At that size it would be one of the largest Hong Kong debuts by a Chinese embodied-intelligence (具身智能) company, and the sharpest test yet of whether public investors will pay private-market prices for a business built on parcel handling.
The 2026 financing run
Robot Era was founded in August 2023, spun out of Tsinghua University's Institute for Interdisciplinary Information Sciences (交叉信息研究院), with Tsinghua holding a little over 6% through its technology-transfer arm. Founder Chen Jianyu (陈建宇), born in 1992, did his PhD at Berkeley before returning to Tsinghua as a professor.
The funding compressed the sector's whole arc into three years:
- A near-RMB 300 million (≈US$41.7M) Pre-A in 2024, then a near-RMB 500 million (≈US$69M) A round in July 2025 and a near-RMB 1 billion (≈US$139M) A+ in November 2025 with Geely Capital and Beijing's AI fund.
- A RMB 1 billion (≈US$139M) strategic round in March 2026 at a valuation above RMB 10 billion (≈US$1.39B), with Samsung and Singtel investing for the first time.
- More than US$200 million (≈RMB 1.44B) on 27 April 2026, led by SF Express (顺丰), and a further RMB 1 billion on 6 July led by Chengtong Fund (诚通基金), a state capital operator under the SASAC, with provincial vehicles from Jiangxi, Anhui and Hangzhou alongside.
Add it up and Robot Era raised more than RMB 3.4 billion (≈US$472M) in 2026 across three disclosed rounds. More than 20 industrial backers sit on the register, from Alibaba and Lenovo to Geely, Dongfeng, Haier, Samsung and Singtel.
Why the capital came in that order
The sequence matters more than the total. March brought overseas strategics. April brought the customer. July brought the state.
Each layer does something different for a company heading toward an IPO. Samsung and Singtel give the equity story an international frame and a supply-chain relationship. SF Express, which led the April round, is also the anchor customer; its money converts a commercial contract into an ownership stake, which helps when an underwriter has to demonstrate demand. A central SASAC vehicle with provincial funds alongside is the kind of cap-table line that reassures cornerstone conversations in Hong Kong.
What is missing is telling: no sovereign wealth fund wrote the largest cheque. The paper sits with parties that have an operational interest in the technology.
The business behind the pitch
Robot Era's commercial story rests on two logistics customers: SF Express and China Post (中国邮政). The company says its robots work in more than ten logistics centres across five provinces, running around the clock, handling 1,200-plus parcels an hour in some configurations and reaching roughly 85% of human efficiency on tasks like label flipping.
It also says it began thousand-unit deliveries in the second quarter of 2026 and expects to ship more than 1,000 units this year, up from dozens in 2025 — a claim that would put it among the few Chinese humanoid (人形机器人) makers with four-digit annual shipments.
The technology claim is a world-model-centric stack: Robot Era published a world-action model called PAD in September 2024, before the debate went mainstream, and its end-to-end ERA-42 model drives full-body manipulation. It also sells dexterous hands (灵巧手) to outside buyers, including global robotics firms and research labs — revenue that does not depend on its own fleet.
Component sales repeat and diversify across customers, but they carry hardware margins and expose the business to the price war that has hit every Chinese robotics supplier.
The question an IPO would have to answer
Public investors will ask what private ones deferred: what does a sorting robot earn per hour, and who pays?
Logistics is the most rational first market for humanoids in China. Sorting centres are semi-structured, volumes are predictable, labour turnover is high. But it is also a market with brutal unit economics: express delivery is priced in fractions of a renminbi per parcel. A robot at 85% of human speed only wins if total cost per parcel — depreciation, maintenance, power, supervision — undercuts a worker costing perhaps RMB 6,000–8,000 a month in a low-tier city.
Two things could tilt that arithmetic. If a machine runs 20-plus hours a day, capital cost per parcel falls fast; three shifts amortise three times quicker than one. And if the same robot can move from parcels to 3C assembly without new hardware, the market widens; partnerships with Geely, Haier and Lenovo point that way, though disclosed volumes are thin.
The risk is concentration. Two customers account for the visible deployments, and a prospectus would have to disclose how much revenue each provides — and whether those contracts are priced to make money or to prove a point.
What the listing would mean
Unitree (宇树科技) has already priced a STAR Market IPO. LimX Dynamics (逐际动力) closed a pre-IPO round in November at about a US$2.2 billion valuation. AgiBot (智元机器人) has been moving toward Hong Kong. Add Robot Era and the sector has gone from raising private capital to competing for public money.
That changes behaviour. In a listing process, a company stops optimising for the next round and starts optimising for disclosed revenue and gross margin, which usually means fewer bespoke deployments and more standard products. It also means the valuations set in 2026 get marked against daily trading prices.
Honest limitations
The IPO size and the adviser discussions come from a Bloomberg report relayed by Chinese tech media; Robot Era has not confirmed a filing, venue, size or timetable, and reported plans can lapse. Round sizes, investor names and the valuation above RMB 10 billion come from Chinese financial media reports of company announcements, not audited filings. Operating metrics — 10-plus logistics centres, 1,200 parcels an hour, 85% of human efficiency, thousand-unit deliveries — are company statements, not independently verified figures. Nothing here is investment advice.
What readers can do now
- Read the customer concentration note first. When a prospectus appears, the revenue split between SF Express, China Post and everyone else will show whether this is a platform or a project business.
- Separate component sales from fleet revenue. Dexterous-hand sales to other robot makers are the most diversifiable line; if that segment grows, the company is less hostage to one sorting contract.
- Compare the listing price with LimX and Unitree, not the last private round. The private mark was set by investors with procurement interests; the public mark by funds that care about cash flow.
