Dong Hui, now 39, of Henan province, suffered a car accident about six years ago that left him unable to move his hands. In November 2024 he became one of the first people in China to receive an invasive brain-computer interface, a coin-sized device called NEO placed on the protective membrane covering his brain. Eleven months later, on the ninth day of a training programme in which he thought his way through a soft robotic glove, his right hand closed around a ball without the glove at all. He has since written his own name, the date, and the words "thank you."
"On the ninth day of my training, my right hand successfully grabbed a ball without the glove. That was a miraculous moment."
Dong Hui told reporters the implant had restored movement in his right hand for the first time since his accident, and said he hoped the technology would eventually reach the many thousands of other spinal injury patients in China still waiting for treatment.
Dong Hui's hard-won victory sits at the centre of a story bigger than any one patient. On 13 March 2026, China's National Medical Products Administration (NMPA) granted commercial approval to NEO, the first invasive BCI sold as a medical product rather than confined to clinical trials [1][2]. The approval placed China, not Neuralink or any other American competitor, at the front of a field that has spent two decades promising to translate thought into action.
A different kind of brain implant
NEO comes from Neuracle Medical Technology, a Shanghai-based firm founded in 2011, and researchers at Tsinghua University in Beijing, where its founder, Xu Honglai, earned his PhD in biomedical engineering [3]. The device is built on a different bet from the rest of the field.
Its eight sensors sit on the dura mater, the tough outer membrane wrapping the brain, with a separate transmitter fixed to the skull. Together they read the electrical signals produced when a patient thinks about moving, decoded into commands for a soft robotic glove worn during roughly 2.5 hours of daily training [1]. The NMPA's approved indication is narrow: adults aged 18 to 60 with quadriplegia from spinal cord injury who still retain some residual function in their arms.
The trade-off is explicit. Avinash Singh of the University of Technology Sydney argues that placing sensors on the dura, not in cortical tissue, lowers the risk of haemorrhage, glial scarring, and signal degradation [1]. The South China Morning Post called it a Confucian "doctrine of the mean", a middle path between invasive precision and non-invasive safety, each with its own clinical costs [4]. Wang Shouyan of Fudan University in Shanghai called the approval hugely important, because it meant BCIs were ready for large-scale manufacturing and clinical use in China [1]. Neuralink's N1 implant, by contrast, carries 3,072 electrodes across 96 threads that penetrate the cortex [5].
How Beijing leapfrogged Silicon Valley
The contrast with the United States is sharp. Neuralink had enrolled roughly 21 participants in its PRIME trial by January 2026 across the US, UK, Canada, and the UAE, and is targeting Phase 3 trials later this year, a premarket approval submission in 2027, and commercial availability in 2028 [2][5]. Synchron, an Australian-American company whose electrode is delivered through a blood vessel, is preparing its own pivotal trial in 2026 [7]. No BCI of any kind has been approved for commercial sale in the US.
Beijing made it a priority. In July 2025, seven Chinese ministries published a plan targeting breakthroughs in core BCI technologies by 2027 and the creation of "two or three" globally leading BCI players by 2030, while the country's latest five-year plan, published the same day as Neuracle's approval, lists BCIs as one of six priority industries [1][8]. Hospitals in Beijing and Shenzhen opened dedicated BCI wards in the first half of 2025, and days after the NMPA's decision, NEO was added to China's national health insurance system [1][8].
Neuracle shows how the system works: the company has run in the order of 36 clinical trials using NEO since October 2023, 32 of them in the first months of 2025 [1]. The Chinese Institute for Brain Research's Beinao-1 system, which completed five human implants in 2025 and decoded Chinese-language speech for an ALS patient with speech impairment, could see domestic approval as early as 2028 [1][9].
The "race" framing is contested by the researchers closest to the technology. Nick Ramsey of Radboud University Nijmegen cautions that a race implies an endpoint, but it is hard to say where that is for BCIs [1]. Meicen Sun of the University of Illinois Urbana-Champaign argues the two countries are running on different tracks, and that being exceptional and being accessible are two diametrically opposed definitions of winning [1]. The undisputed fact is that China holds the only commercial approval for an invasive BCI anywhere in the world.
Privacy, sovereignty, and the brain as a platform
The flip side of putting paralysed patients on a BCI is putting brains on a network. A July 2025 Yale study in Neuroethics found most leading BCIs lack encryption for the wireless data stream, as the power budget inside the skull is too tight to run strong cryptography [10]. Tyler Schroder warned that a breach in standardised BCI systems could affect millions at once, enabling mass manipulation of neural data or cognitive impairment [10].
Schroder and the centre's director, Luciano Floridi, recommended non-surgical methods for updating devices, strong authentication for software changes, and end-to-end encryption of the data moving in and out of a patient's brain [10]. In the US, BCIs are already classified as Class III implantable medical devices, the most stringent category, but the software on them is loosely regulated, a gap the Yale team argues must be closed before the next generation ships. Whether China's NMPA-regulated devices face an analogous software-oversight gap is not yet publicly documented. Singh put it bluntly: there is no comparable national-level ambition or coordinated map elsewhere in the world at the moment [1]. When a country treats the brain as the next platform, ownership of the data inside stops being a privacy footnote and starts being a sovereignty question.
Dong Hui completed an eleven-month rehabilitation programme. He told MIT Technology Review the device will help not only him but also the thousands of other patients in China living with paralysis, bringing them hope [1]. The NMPA's decision makes NEO the only invasive BCI in routine commercial use, and him the most public face of an early clinical achievement. The strategic significance of that achievement, and the choices the rest of the field makes over the next five years, will quietly answer what "winning" the era means.