--- title: "Three Implant-Grade Approvals in Twelve Months: China's BCI Clears the Regulatory Bar It Could Not Clear Before" date: 2026-09-07 category: Brain–Computer Interface site: NeuroAI canonical: https://neuroai.site/a/na-bci-class-three-approvals language: en --- # Three Implant-Grade Approvals in Twelve Months: China's BCI Clears the Regulatory Bar It Could Not Clear Before > Class III is China's highest medical device risk class — long-term implant, strictest review. Since early 2026, multiple BCI devices have cleared it, shifting the industry's question from 'is this possible?' to 'how fast can it reach patients? ## Key takeaways - On **4 September 2026**, Nanjing Shanhai Medical's combined stimulation-recording electrode received a national Class III medical device registration — the first domestically approved electrode that both records (Spike/LFP) and stimulates through the same dual-macro tip. - In **March 2026**, NeuroXess's implantable system became the world's first implantable BCI medical device to be approved for market. Shanhai's earlier "Mingtong" device won China's first BCI Class III certificate in September 2025. - Class III means the device may be implanted in the human body long-term. Most BCI products until now held Class II — external, short-duration, lower risk. - The queue is forming: on the same day, Shuli Innovation formally began STAR Market IPO tutoring. **No company listed in mainland China has BCI as its core business**, and at least four are in the pipeline. ## What one electrode changed Have you ever sat with an elderly relative in a neurology outpatient clinic? The hand shakes too badly to hold a bowl. When the medication wears off they freeze on the sofa. The doctor looks at the scans, says surgery could be considered — and then adds that assessment takes time, and there is a queue. Behind that sentence is the most honest description of where brain-computer interfaces actually stand in 2026. Deep brain stimulation (DBS) surgery works like this: a surgeon threads an electrode deep into the brain, and must simultaneously do two jobs. First, read neural signals to confirm "is this the right spot?" Second, apply stimulation to check "did we hit it?" Historically those were two separate instruments, swapped back and forth. Every swap is a risk and a delay. The electrode approved on 4 September does both through one device. One fewer instrument change means one more margin of safety in the surgeon's hands. ## Why three words matter: Class III China classifies medical devices into three tiers. Class III is the top: highest risk, longest review, and the only class that permits long-term implantation in the human body. For years, BCI products mostly carried Class II — external use, short duration, manageable risk. That began changing in 2026: - **NeuroXess**: implantable system approved in March, the first of its kind anywhere in the world to reach market - **Shanhai Medical**: one Class III certificate in September 2025 and another in September 2026 - **Lepu Medical Electronics, Jingyu Medical** and others have also secured approvals Approvals arriving in batches while companies queue for listings — those two things happening at once tells you something specific. The question of *whether it can be done* has essentially been answered. What remains is *how fast it can spread*. ## Three constraints on how long the queue will be Approval does not automatically produce access. Three things sit in between. **There are not enough surgeons.** These procedures depend heavily on the operator's experience. The number of hospitals where they are performed independently is concentrated in a handful of top-tier centres. **Cost and reach.** The device, the consumables, the surgery itself, and the post-operative programming that goes on for years afterwards add up to a long bill. **The boundary of indications.** These devices treat defined conditions. They do not treat "I would like better memory." That boundary is a feature, not a limitation — but it is widely misrepresented in marketing. Technology moves faster than hospitals, and hospitals move faster than insurance. The gap between them is what an ordinary family feels as "how much longer?" ## What is actually worth doing First, if someone in your family has Parkinson's disease or essential tremor, ask once whether they are a candidate for assessment. A large number of people are not ineligible — they have simply never been referred. Second, learn to look for the Class III mark. Many products on the market call themselves "brain-computer interface," especially in the rehabilitation category. Certified and uncertified are two entirely different things. Third, resist the science-fiction frame. What this technology does today is reconnect a wire that broke inside a body, so that a hand obeys again. That is already a large thing. The real progress here is not humans acquiring superpowers. It is an elderly person who could not hold a bowl steadily being able to lift it to their mouth again. ## Honest limitations Class III approval establishes safety and efficacy for a defined indication under Chinese regulatory review; it is not a global endorsement, and approvals by other regulators may follow on different timelines. The statement that "no mainland-listed company has BCI as its core business" reflects the market as of September 2026 and will likely change. IPO tutoring is an early procedural step, not a filing. *Sources: National Medical Products Administration registration announcements (4 September 2026, March 2026); company disclosures; STAR Market tutoring filings. Information only — not investment or medical advice.* --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-bci-class-three-approvals Free to quote with attribution and a link to the original.