If you follow brain–computer interfaces (BCIs) in English-language media, the story is basically one company: Neuralink, founded by Elon Musk. So it surprised many readers when, in early 2026, a Chinese implantable BCI became the first in the world to clear commercial approval — ahead of Neuralink's still-small human trial.
The headline "China overtakes Neuralink" is tidy but incomplete. The two programs are solving different problems, on different regulatory tracks, with different definitions of success. Here is what is actually different.
Key takeaways
- Neuralink (US) builds the N1 implant and the R1 surgical robot, running the PRIME study; as of 2026 it remained in a roughly 9-subject trial phase.
- NEO, from Chinese firm BrainRobotics (博睿康), received China's Class III medical-device certification from the NMPA (National Medical Products Administration, 国家药监局) in March 2026 — described by MIT Technology Review and others as the first implantable BCI to complete trials and win commercial approval anywhere.
- The approved NEO use is narrow: helping spinal-cord-injury patients compensate for lost hand movement.
- The real race may not be implantable at all — non-invasive, AI-driven BCI aimed at chronic pain and rehabilitation could reach far more people first.
The headline gap: approval, not ability
Neuralink has implanted its N1 device in a small number of patients and published compelling demos — controlling a cursor, playing a game, moving a robotic arm with intention. Those are real. But "approved for commercial use" is a different bar from "implanted in a research study."
NEO cleared that bar first. Its NMPA Class III certification means a regulator judged the device safe and effective for a defined medical indication, not merely allowable in a trial. That is why observers called it the first implantable BCI with commercial approval. Neuralink's path runs through the US FDA and its own controlled trial expansion.
What each one is built for
Neuralink's public narrative is broad and science-fiction-adjacent: restoring autonomy to paralyzed people, and eventually a generalized brain interface. Its current approved research use centers on motor-intent control for people with quadriplegia.
NEO's approved use is deliberately narrow: compensating for lost hand movement in spinal-cord-injury patients. In an August 2026 implantation at a prefecture-level hospital in Weihai (威海), a patient who could not grip for years regained a connection between the brain's "I want to hold the cup" command and the hand. Narrow, but it is already a deployed, approved therapy rather than a trial milestone.
The regulatory contrast
The difference is partly structural. The US system prizes a slow, conservative, per-device trial expansion — safe, but slow to reach patients. China's NMPA moved NEO through Class III certification and, from the 15th Five-Year Plan (十五五, 2026–2030) onward, wrote BCI into national policy, with hospitals opening BCI wards and connecting to clinical use.
Neither approach is "wrong." They optimize for different things: the US for procedural caution, China for speed-to-patient once a use case is bounded. What changed in 2026 is that the approval milestone landed in China first.
The non-invasive wildcard
Both implantable programs are capped by the same physics: every patient is another operation. A rival route skips surgery entirely. Non-invasive ultrasound BCI, flagged by CAS academician Zheng Hairong (郑海荣), reads and writes neural signals across the whole brain without cutting, aimed at China's more than 300 million chronic-pain sufferers, plus stroke rehab and Alzheimer's intervention. Several tertiary hospitals have launched multi-center validation.
If "winning" means reaching the most patients, a non-invasive device that needs no operating room may outrun any implant — including NEO and Neuralink alike.
What "winning" actually means
A commercial license is not automatically a business. NEO's indication is real medicine but a limited pool, with a high per-device price and heavy reliance on hospitals and skilled surgeons. Neuralink's eventual market depends on how far its indication expands. The investor's question — how does one surgery become sustainable cash flow? — applies to both.
The deeper shift is where the moat sits. As AI pushes decoding accuracy from roughly 60% to above 90%, the bottleneck moves from hardware to data and scenario. Whoever holds real clinical data and sits closest to doctors and patients holds the entrance — whether the device is invasive or not.
Honest limitations
Approval-status and comparison claims reflect company disclosures and reporting by MIT Technology Review, not an independent regulatory audit reproduced here. The "world first" characterization reflects how observers describe the approval and has not been benchmarked against every global filing. The 60%–90% decoding-accuracy figure is attributed to researchers and industry estimates in coverage, not a single peer-reviewed source. Neuralink's trial-subject count reflects publicly reported figures and may change as trials expand.
Sources
Company disclosures from Neuralink and BrainRobotics; MIT Technology Review's characterization of NEO's approval status; remarks by CAS academician Zheng Hairong (郑海荣); and 2026 reporting on BCI clinical deployment and investment.