The company was registered in February. By November it held RMB 794 million of other people's money.
Shanghai Zhudong Technology (主动科技) has not published a single human implant. No trial date, no named hospital, no patient data, no disclosed valuation. What it has is a 1,024-channel brain-computer interface (脑机接口) system that cleared a pre-clinical safety review, and investors — most of them state-backed — willing to fund the distance between those two facts.
On November 4 the company said it closed an angel-plus round (天使+轮) of RMB 464 million (≈ US$69M / HK$542M). That follows an RMB 330 million (≈ US$49M / HK$386M) angel round announced on August 3. Three months, two rounds, RMB 794 million (≈ US$118M / HK$928M).
Key takeaways
- Zhudong closed RMB 464 million on November 4, 2026, about three months after a RMB 330 million angel round. Together, RMB 794 million (≈ US$118M / HK$928M).
- The company was founded in February 2026 and incubated by Lingang Laboratory (临港实验室). The 1,024-channel system's type inspection report (型式检验报告) dates from November 2025 and was obtained by the incubating team, before the company legally existed.
- As of the announcement, Zhudong had disclosed no human implant, no trial start, no lead hospital, no enrollment figures and no post-money valuation.
- Two weeks earlier, the medical device regulator finalized trial rules for implantable grasp-restoration systems — and excluded intracortical implants, the category Zhudong is building.
- China's BCI sector drew more than RMB 7 billion (≈ US$1.04B) in the first half of 2026, more than double all of 2025, per ITjuzi figures cited by Economic Information Daily (经济参考报).
What the company actually sold
The pitch rests on one number: 1,024 channels — the width of the pipe through which a device listens to neurons. Zhudong says it owns the full stack: flexible electrodes, a dedicated chip, encoding and decoding algorithms, system integration.
Its lead product, Tianshu (天枢), targets motor and speech impairment from stroke, spinal cord injury and ALS. Two further lines, vision and emotion, remain research directions with no published names, specifications or timetables.
The investor mix is unusual for a nine-month-old company:
- Co-leads: Gaorong Capital (高榕创投), Shanghai Guotou Xiandao (国投先导), Shanghai Science and Technology Innovation Group (上海科创投).
- National platform: China Internet Investment Fund (中网投), described by the company as the round's only national-level investor.
- Corporate and local money: Guangzhou Pharmaceutical Capital (广药资本), Guotai Haitong (国泰海通), Pudong Venture Capital (浦东创投), Minhang Investment (闵金投), plus earlier backers Daotong Capital (道彤资本) and Huada Songhe (华大松禾).
Chief executive Guojia Wu (吴国佳), previously at Boston Scientific and an executive director at CardioFlow (心通医疗), said proceeds would fund research and three clinical programs.
Why the state arrived at the angel stage
Public money in Chinese deep tech normally waits for a registration certificate. Here it arrived before the first incision.
Geography helps. Shanghai has been building a BCI cluster around Zhangjiang and the Brain Intelligence World (脑智天地) park, with Lingang Laboratory as an anchor. A lab spinoff staffed by people local funds already know is an easier call than a stranger with a deck.
Policy explains more. Brain-computer interfaces were named in the 2026 Government Work Report and in the 15th Five-Year Plan as a future industry. Once a category reaches that list, local governments have an incentive to seed it early rather than bid later.
The hardware is real too: a 2,500-square-metre office and GMP cleanroom are built, with a 10,000-square-metre headquarters planned by end-2026. Investors are funding a factory with nothing yet to make.
The rulebook landed two weeks earlier — and points elsewhere
On October 22, the device evaluation centre of the National Medical Products Administration (国家药监局) issued formal clinical trial review points for implantable BCI hand-grasp systems. A July 31 draft became binding.
The specifics matter:
- Feasibility trials precede confirmatory trials, with no fewer than four cases in the feasibility stage.
- Post-market follow-up should run at least two years after implantation.
- The target population is 18 to 60, with C2–C6 cervical spinal cord injury, AIS grade A to C, at least 50% grade A.
- It covers low-channel designs — citing under 64 channels as the reference — implanted outside the dura.
Then comes the line that should stop a reader. The guidance explicitly does not apply to implants placed under the dura, inside the cortex, deep in the brain, or inside blood vessels. Zhudong's 1,024-channel intracortical system sits outside China's first formal clinical rulebook for this device class.
Not a scandal: the regulator wrote the rule for the product that got there first; the record early-stage round went to the product that has not.
Neuracle (博睿康) won approval from the National Medical Products Administration on March 13, 2026 for NEO — an epidural device with eight electrodes and a coin-sized implant, the first implantable BCI device approved in China. It had run 36 procedures, four feasibility and 32 confirmatory, and was on Shanghai's medical insurance consumables catalogue within two weeks. That is what a de-risked asset looks like.
Two clocks
Every deep-tech bet runs on two clocks: how long a company can pay engineers, run studies and keep a cleanroom certified; and how long until a person moves a hand because of this device, and whether the effect holds.
Zhudong has bought a great deal of the first clock and not started spending the second. It said in August it planned to begin trials with the 1,024-channel system; by November it still had no date.
Channel count is not durability. A wider pipe carries more signal, but the electrode decides whether it is still there in year three.
None of this makes the bet irrational. It makes it early. Investors are buying the option to still be at the table when the evidence clock starts.
Honest limitations
- Zhudong has not disclosed its post-money valuation, so no return arithmetic is possible. The "record" claim is the company's own, based on announced deals, and is scoped to angel and early-stage rounds.
- A type inspection report (型式检验) is a design and performance check, not a registration certificate. Neither the testing laboratory nor the report number appears in public materials, and we could not verify either.
- Absence of disclosure is not absence of activity. We report only what was public as of November 4.
- Primary sources are Chinese-language: the company announcement, Shanghai Securities News (上海证券报), Wallstreetcn (华尔街见闻) and the regulator's notice. No English wire story on this round was found.
- Conversions use the November 16, 2026 central parity rate (RMB 6.7141 per US dollar, RMB 0.8558 per Hong Kong dollar), rounded. This article does not assess the technology's clinical merit.
What readers can do now
- Track three dates, not the headlines. A trial start, a named lead hospital and a first published implant are the only disclosures that change the risk profile. Until then, treat any BCI valuation as pre-evidence.
- Separate the two documents. A type inspection report and a registration certificate are different objects. If a pitch deck blurs them, ask for the report number and the testing lab.
- Read the October 22 guidance yourself. If a company's device falls outside its scope, its regulatory path is longer than the sector narrative implies; its money buys time, not certainty.
The money arrived in ninety-three days. The evidence will not.
