Most brain implants stay wherever the surgeon puts them, and reaching another part of the brain often means another hole in the skull. Junrou Technology, a Chinese startup, is building electrodes that can be moved after they go in, steered through the head by an external magnet from a single small opening. Its angel-plus round of nearly RMB 100 million, announced on 2 November, was led by Shunwei Capital, the fund co-founded by Xiaomi’s Lei Jun, with GL Ventures and humanoid robot maker AgiBot among the other investors.
The round
Shunwei led the round. GL Ventures, the early-stage arm of Hillhouse, followed alongside AgiBot and Nansha Qisheng. The company says the money will go towards finalising the manufacturing process for its electrodes, building a GMP production line, animal studies and its neural-signal algorithms.
Junrou has closed three rounds since it was founded in 2025, taking total funding to about RMB 100 million. Jiuhe Venture Capital led the previous round, with the National Innovation Center for Advanced Medical Devices following.
Shunwei already holds a stake in Beijing implant developer Zhiran Medical, where it was named as an existing shareholder in a RMB 300 million round in February.
The technology
Founder Liu Zhiyuan is a researcher at the Chinese Academy of Sciences’ Shenzhen Institutes of Advanced Technology and executive dean of the artificial intelligence research institute at Shenzhen University of Advanced Technology. His partner in the company is Peng Qianbi, a doctoral student he co-supervised. Chinese Academy of Sciences academician Zheng Hairong is its chief academic adviser.
The company is built on NeuroWorm, an electrode published in Nature in September 2025. Liu was one of four corresponding authors on the paper, with Han Fei, Yan Wei and Xu Tiantian. The electrode is a soft, stretchable fibre about 0.2 millimetres wide, made by rolling a flat electrode array into a thread, with a magnetic component at its tip. An external magnetic field pulls the tip forward and turns it, so the fibre can be guided to a target while it records. Liu and the paper’s first-listed author, Xie Ruijie, are named in the paper as inventors on a Chinese patent application.
Junrou’s plan is to pass several of these fibres through one small opening in the skull, which Liu puts at about a centimetre wide, and steer each to a different region of the brain. Its first system has completed part of its animal testing, and the company expects to begin an investigator-initiated clinical study in 2027, starting with epilepsy and depression, conditions that involve several brain regions. Liu says later generations will use the signals they record to find their own way to the most useful sites.
“Imagine opening 30 holes in the skull,” Liu told Waves, 36Kr’s investment publication. “It would end up looking like a honeycomb.” How little damage an implant does, he said, decides whether more people will accept one.
The robot connection
The same flexible sensing materials go into a second product. Junrou has developed a stretchable electronic skin that attaches to dexterous robot hands and robot bodies and senses pressure, texture and the spread of force, a product aimed at humanoid robots.
Liu says Chinese investors have changed since he first discussed the idea with some of them after returning from Singapore in 2019. Back then, he said, investors told him: “What you’re describing, America doesn’t have yet.” Now, he said, they are increasingly willing to back ideas the market has not yet agreed on.
Signal: Why it matters
Electrodes that can be repositioned would let one implant cover several brain regions and follow a disease as it changes, with less surgery each time. Liu argues that safer, less invasive implants will widen the pool of patients willing to have one, and with it the amount of brain data a company can collect. In this round, Lei Jun’s venture fund, Hillhouse’s early-stage arm and a humanoid robot maker all backed the same materials platform, which serves both brain implants and robots.
Not disclosed
The size of each earlier round, the company’s valuation and where it is headquartered have not been published. Junrou has not named the hospital for its planned study or said how many patients it will include. Junrou has reported no human data.
Forward Radar: What to watch
Whether the investigator-initiated study starts in 2027, and at which hospital.
Whether the steering holds up safely in people, which is the claim the whole company rests on.
Whether AgiBot or another humanoid maker puts Junrou’s electronic skin into its robots.
Sources
Funding:
- Junrou Technology announcement, posted by Lighthouse Capital, 2 November 2026 (Chinese)
- 36Kr (Waves): Shunwei, GL Ventures and others invest RMB 100 million in a ‘neural worm’ brain-computer company, posted 10 November 2026 (Chinese)
- Dahe Cube via 10jqka: Lei Jun and Hillhouse invest in a brain-computer interface company, 2 November 2026 (Chinese)
- Beijing Business Today via Sohu: Junrou Technology completes nearly RMB 100 million angel-plus round, 2 November 2026 (Chinese)
- IT Home via Phoenix New Media: Junrou Technology backed by Shunwei Capital and GL Ventures, 2 November 2026 (Chinese)
Research:
Cross-reference to prior InsideBCI coverage: