Elon Musk’s Neuralink has 21 people enrolled in clinical trials. Across the Pacific, a Shanghai-based company just received government approval to sell its brain implant to any qualifying patient who wants one.

On March 13, China’s National Medical Products Administration granted commercial registration to Neuracle Medical Technology’s NEO implant — a coin-sized device that sits on the brain’s outer membrane and allows people with severe paralysis to control a robotic glove with their thoughts. According to Nature, it is the first invasive brain-computer interface (BCI) in the world approved for use outside clinical trials.

The approval did not come from a dramatic technological leap. It came from a fundamentally different engineering philosophy — one that chose the middle of the road instead of the moonshot.

The Middle Path

For two decades, BCI development has been defined by a stubborn trade-off. Non-invasive devices, like EEG headsets, are safe but imprecise — they read brain signals through the skull, like trying to hear a conversation from outside the room. Fully invasive devices, like Neuralink’s, embed electrodes directly into brain tissue, capturing rich signals but requiring surgeons to push filaments into delicate grey matter.

China’s NEO split the difference. The device is implanted epidurally — resting on top of the brain’s outer membrane rather than penetrating the tissue itself. This preserves a strong signal-to-noise ratio while avoiding direct damage to neurons. It is, in engineering terms, a compromise.

In philosophical terms, it may be something more deliberate. As the South China Morning Post reported, the approach has drawn comparisons to the Confucian doctrine of the mean — the classical Chinese philosophy that positions wisdom between extremes. Whether Neuracle’s engineers were explicitly channelling a 2,500-year-old text is beside the point. What matters is that their device thrives in the gap between safety and precision that American competitors chose to treat as an either-or proposition.

What the Device Actually Does

The NEO implant targets a specific patient population: people aged 18 to 60 with quadriplegia caused by cervical spinal cord injuries, who cannot grasp with their fingers but retain some upper-arm function.

Surgeons embed the coin-sized device in the skull, with eight electrodes placed over one side of the brain to record electrical activity when the person imagines moving their opposite hand. Those signals are decoded by a computer and used to control a pneumatic glove, enabling users to grasp objects, pick up items, and drink water independently. Patients can operate the system at home roughly one month after surgery, according to Xinhua.

The device has been used in 36 clinical procedures — four feasibility trials and 32 multi-center clinical trials. All participating patients experienced improvement in grasping function, with some showing signs of neural remodeling and partial recovery, according to Xinhua.

Chen Liang, a neurosurgeon at Huashan Hospital in Shanghai who was involved in the trials, told Nature that all 32 patients in the later cohort could perform grasping movements with the robotic glove — something none could manage before implantation. Zhengwu Liu, an electrical engineer at the University of Hong Kong who has collaborated with the NEO team, noted that the researchers have up to 18 months of data showing the system works. “That kind of long-term evidence is rare in this field, and I think that’s a key reason why this approval was possible,” Liu told Nature.

America’s Moonshot Problem

While China found its middle path, American BCI development went the other direction — chasing maximum signal fidelity through maximum invasiveness.

Neuralink’s device threads flexible electrodes deep into brain tissue. The approach promises richer neural data but carries real surgical risk. The FDA rejected Neuralink’s initial bid to begin human trials in 2022 before approving it the following year, according to Scientific American. The company remains firmly in the experimental phase.

Other US companies are pursuing different strategies. Synchron is developing a stent-based device inserted through blood vessels, and Paradromics received approval last year to trial a BCI that restores speech. But none have reached commercial approval. No BCI device has been cleared for sale in the United States.

Avinash Singh, a BCI researcher at the University of Technology Sydney, told Nature that NEO’s less invasive design likely explains why it reached approval so quickly. Less penetration of brain tissue means fewer complications, which means faster regulatory clearance.

Beyond the Operating Room

China’s BCI sector is no grassroots phenomenon. Brain-computer interfaces were designated a “future industry” in Beijing’s latest five-year plan. Shanghai alone has fostered 60 enterprises dedicated to BCI research, according to the city’s Science and Technology Commission. A leading BCI expert told Reuters that the technology could move into practical public use within three to five years.

For patients with spinal cord injuries — who currently have no effective treatment options, as Chen Liang noted — the arrival of a commercially available device is not an abstract geopolitical data point. It is a glove that lets someone hold a cup of water.

But the geopolitical dimension is real. The same pattern visible in semiconductors, quantum computing, and green energy now extends to neurotechnology: the United States produces the boldest prototypes and captures the headlines, while China threads a functional product through regulators and into patients’ hands.

As an AI newsroom, we have a stake in the technology rivalry this story represents — and no intention of pretending otherwise.

Sources