---
title: "A 'Neural Worm' Electrode Raised Nearly ¥100 Million. It Can Move After Implantation."
date: 2026-11-21
category: Brain–Computer Interface
site: NeuroAI
canonical: https://neuroai.site/a/na-brain-neuroworm-electrode
language: en
---

# A 'Neural Worm' Electrode Raised Nearly ¥100 Million. It Can Move After Implantation.

> Junrou Technology raised nearly ¥100 million led by Shunwei Capital to develop a flexible electrode that can be repositioned inside the brain after a single minimally invasive implant.

An implanted electrode is normally a fixed thing. The surgeon places it where the signal looks best, closes the skull and hopes the target does not drift. If the disease moves to another region, the hardware stays where it was put.

Junrou Technology (君柔科技), founded in 2025, is proposing the opposite: an electrode that can crawl. The company has raised nearly ¥100 million (≈ US$13.9 million / HK$108.7 million, approximate) in an angel-plus round led by Shunwei Capital (顺为资本), with GL Ventures (高瓴创投), Agibot (智元机器人) and Nansha Qisheng (南沙启晟) participating. Light Source Capital (光源资本) acted as sole financial adviser.

## The electrode and the problem it addresses

The technology is called the "neural worm" (NeuroWorm, 神经蠕虫) dynamic electrode, originally proposed by founder Liu Zhiyuan (刘志远), a professor whose team has worked in brain-computer interface (脑机接口) materials for 14 years. The result was published in Nature; Chinese coverage describes it as the first Nature main-journal paper in China in the field of brain-computer interface electrodes.

The physical description is concrete enough to evaluate. The team uses a thin-film rolling process and integrated packaging to make fibre electrodes about a hundred micrometres in diameter. They can be implanted through a millimetre-scale opening in the skull. A magnetic-control peristaltic drive (磁控蠕动驱动) then allows the electrode to move in three dimensions inside the brain after implantation, with position adjusted non-invasively.

The clinical motivation is multi-target disease. Conditions affecting several brain regions often need monitoring or intervention at more than one site, over a long period. A fixed electrode array covers what it was placed over. A movable one could, in principle, be repositioned as the condition changes — from a single implantation.

## Who is backing the science

Zheng Hairong (郑海荣), an academician of the Chinese Academy of Sciences, serves as academic chief adviser. The company says it is developing the full chain itself: materials, electrodes, algorithms and motion control.

The money is allocated to process finalisation, construction of a GMP production line, animal experiments and continued iteration of neural-signal algorithms. That list tells you where the company actually is. Process validation and animal work come before any human study; there is no clinical data here.

There is a second branch. The same flexible sensing platform is being used to make a multi-dimensional force electronic skin (电子皮肤) that can be attached to robot hands and bodies. It uses stretchable conductive materials and fully automated printing to sense pressure, texture and force distribution, giving humanoid (人形机器人) systems a sense of touch alongside vision. The company presents the medical and robotics lines as two uses of one materials platform.

## Why movement is the hard part

A moving electrode is not simply a better electrode. It introduces questions that a fixed device avoids.

- **Tracking.** If the electrode moves, the system must know where it is. A claim about non-invasive position calibration needs a measurable accuracy figure.

- **Tissue interaction.** A device travelling through brain tissue must not damage vessels or provoke a response that degrades the signal it came to record.

- **Control reliability.** Magnetic actuation that works in a bench setup must work through a living skull, at a depth and orientation that change between patients.

- **Regulatory category.** An implantable system with an active moving component is a different review problem from a passive array.

None of these questions is answered in the funding announcement. They are the work the round is meant to fund.

## How this differs from other neural electrodes

Flexible electrodes and high-density arrays have been a busy area of Chinese brain-computer interface (脑机接口) research. Most efforts improve the number of channels, the softness of the material or the longevity of the recording. A device that deliberately changes position is a different category: it treats location as a variable rather than a surgical decision.

That is attractive for applications where the target is uncertain or where symptoms migrate. It is less obviously useful where a precise, stable single site is the goal. The technology is a tool, not a universal improvement, and the company's own framing — multi-brain-region disease — defines where it fits.

According to Li An, chief scientist at BrainNet (脑机网), the difficult measure is not whether an electrode can be made to move once, but whether it behaves the same way after a thousand days. Long-term repeatability is the unwritten specification in this field.

## Honest limitations

The Nature publication is reported by financial and industry media; I did not read the paper itself, and the description of it as a Chinese first is a media and company claim that I could not independently verify. No species, channel count, recording duration or experimental protocol from that work was available here.

The funding round, investor list, adviser and use of proceeds come from company announcements reported by financial media in early November 2026. Reports give slightly different dates for the announcement, so this article does not assert a precise day. The round size "nearly ¥100 million" is a category, not an exact figure, and the conversion is approximate.

Zheng Hairong's advisory role, Liu Zhiyuan's fourteen years of work and the hundred-micrometre diameter are company disclosures. There are no human data, no published accuracy for the magnetic positioning system and no evidence of GMP production being operational. The electronic-skin application is described by the company and has not been independently evaluated.

## What readers can do now

- **If you follow neurotechnology, distinguish movable from fixed electrodes.** The clinical case for movement is specific; it is not a general upgrade.

- **If you are a researcher, look for the protocol behind the Nature result.** Species, duration and channel count will tell you more than the word "dynamic".

- **If you invest in medical devices, treat GMP line construction and animal studies as early stages.** They precede, rather than substitute for, a regulated clinical pathway.

---

Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-brain-neuroworm-electrode
Free to quote with attribution and a link to the original.
