The more you follow news about the international BCI race, the harder it becomes to tell which country is actually ahead. In March 2026, headlines kept announcing that "China has overtaken Neuralink," because China's National Medical Products Administration (NMPA) granted the world's first commercial approval to NEO, an invasive BCI developed by Shanghai-based Neuracle Medical Technology (博睿康医療科技). Yet the clearance Precision Neuroscience received in the US covers only temporary use, for under 30 days, of a passive cortical electrode array with no software and no wireless capability. The yardstick for "who is ahead" differs enormously from country to country. The absence of a shared yardstick is the real problem this approval drama reveals.
NEO's approval has gone as far as an insurance code for a permanently implanted BCI
On March 13, 2026, the NMPA issued a Class III medical device registration certificate to NEO, an implantable BCI developed by Neuracle Medical Technology (博睿康医療科技, based in Zhangjiang, Pudong New Area, Shanghai). Its formal name is the "Implantable Brain-Computer Interface Hand Motor Function Compensation System," and it is limited to patients aged 18 to 60 who have quadriplegia from cervical spinal cord injury but retain some upper-arm function. The technology builds on research originating at Tsinghua University and uses a semi-invasive design: a coin-sized disc placed beneath the skull and above the dura. Because it does not penetrate brain tissue, its structure is fundamentally different from fully invasive BCIs that insert electrodes into the brain.
Just two days after the approval, on March 15, China's National Healthcare Security Administration (NHSA) set up a dedicated channel itself and completed issuance of a medical insurance code. The code is the "entry ticket" that lets medical institutions treat NEO as an insurance-claimable item, and the agency itself acknowledged that "whether the real-world clinical benefit is worth paying for through insurance, including hospitals' willingness to adopt it, still involves real-world bargaining." It is important to note that issuing the code does not mean reimbursement is settled. Even so, no case of a BCI in any other country has reached this point just two days after approval.
NeuroXess, also based in Shanghai, is running trials of a fully implanted, fully wireless BCI, but as of July 2026 it has not obtained commercial approval and is a separate company at a different stage from NEO. Lumping everything together as "China's BCIs" risks confusing approved products with those still in trials.
The gap between China's Class III approval and America's 30-day diagnostic clearance
Precision Neuroscience's "Layer 7-T" received 510(k) clearance (a streamlined premarket notification; K242618) from the FDA on March 30, 2025. However, it is a passive cortical electrode array with no software or wireless communication, a simple device that operates on power from an external FDA-cleared EEG system. Its use is limited to temporary implantation of under 30 days to record, monitor, and stimulate electrical signals on the brain surface, in settings that support physicians' diagnostic work such as intraoperative brain function mapping. It belongs to an entirely different product category from a "BCI as a system" like NEO, which has decoding software and wireless transmission and converts a patient's movement intent into the operation of an external device.
As of mid-2026, no implantable BCI intended to restore motor function has received full FDA approval (PMA) in the US. Neuralink is running multiple clinical trials but has not crossed that line. The FDA is cautious about permanent implants meant to compensate for motor function because it requires accumulated long-term safety data, including on infection and electrode biocompatibility.
Synchron, which develops the endovascular, non-craniotomy BCI "Stentrode" placed via catheter, sits between the two. On November 6, 2025, it raised a $200 million Series D (about ¥32.5 billion at ¥162.40 to the dollar), bringing total funding to $345 million. It co-developed a Bluetooth protocol (BCI-HID) for connecting with Apple's iPhone, iPad, and Vision Pro, and is moving toward a US pivotal trial, but it has not yet reached approval itself.
Neuracle, which has gone as far as a permanent implant and an insurance code, and the Chinese government, which gains a strategic advantage at the national level, are the first to benefit from this asymmetry in regulatory categories. Neuralink and Precision Neuroscience, busy with regulatory compliance, lose time, while US patients, in a market racing ahead without a binding federal neural data protection law, are positioned to bear the strain (some states, including Colorado and California, have independently added neural data to the data they protect). China's insurance code system also puts it in the position of being first in the world to accumulate cost-effectiveness data on permanently implanted BCIs, which could become a reference point when other countries' insurers consider reimbursement models.
Market forecasts vary by research firm, but Grand View Research projects $6.2 billion in 2030, Research and Markets $4.66 billion, and Allied Market Research $5.463 billion, all converging at roughly $5 billion to $6.5 billion. At the same exchange rate, $6.2 billion equals about ¥1.0069 trillion. Compared with the AI and semiconductor markets, BCI remains a niche emerging field. Even so, inside that small market, a contest over who sets the standard has already surfaced in the form of differences in approval categories.
How the semi-invasive design reads hand movements with eight sensors
At the core of NEO is a disc-shaped device about 2 to 3 centimeters in diameter, implanted beneath the skull and above the dura. Eight sensors pick up electrical signals from the motor cortex of the cerebral cortex and detect the pattern of neural activity that arises when a patient intends to "grasp a hand" or "extend an arm." The signals are sent wirelessly from inside the body to external decoding software, which converts them into the type of intended movement and then operates an external device such as a pneumatic glove.
The decoding software learns in advance the neural activity patterns that appear when a patient intends a specific movement, then matches incoming real-time signals against those patterns and classifies them into discrete movement commands such as "grasp" and "release." It does not reproduce continuous movement itself. Even so, classification into a limited set of movement commands alone is practical enough to compensate for everyday grasping.
BCIs fall roughly into four levels of invasiveness: fully invasive types that insert electrodes into brain tissue itself (such as Neuralink); semi-invasive types that touch the brain surface but do not penetrate tissue (such as NEO and Layer 7-T); non-invasive types that pick up signals from outside the scalp (such as EEG caps); and endovascular types that pass a catheter through blood vessels to veins near the brain (Stentrode). The higher the invasiveness, the higher the signal precision, but so too the risks of infection and bleeding associated with craniotomy. The semi-invasive design is a realistic middle ground between precision and risk, and this design decision may have been one factor behind the outcome of an insurance code for a permanent implant.
The neural data protection gap: Chile's precedent from five years ago and the stalled US inquiry
What a BCI reads is extremely personal neural data on a patient's movement intent. Legal systems to protect this kind of data are developed to very different degrees from country to country. In 2021, Chile set a precedent as the first country in the world to amend its constitution to cover the neuro field (Ley N°21.383). The amended text states that "scientific and technological development shall serve people and be carried out with respect for life and physical and mental integrity," obligating the state to provide special legal protection for brain activity itself and the information derived from it. A separate neurorights protection bill, which would specify in concrete terms matters such as a ban on sales, free will, and equal access to cognitive augmentation technologies, remains under deliberation as a follow-up to the constitutional amendment.
The current US situation contrasts with Chile's. On September 29, 2025, Senator Chuck Schumer, together with Senators Cantwell and Markey, introduced the "MIND Act (S.2925)." But the bill does not regulate neural data itself. It merely directs the Federal Trade Commission (FTC) to investigate how neural data is handled and analyze regulatory gaps, making it essentially a bill that orders a study.
China also has no protection law dedicated to neural data. The "Implementation Opinions on Promoting the Innovative Development of the Brain-Computer Interface Industry," announced by seven ministries on July 23, 2025, sets industrial goals of breakthroughs in core technologies by 2027 and the nurturing of two to three global companies by 2030, and also mentions building a data governance framework and "preventing leaks of brain privacy." But it does not address implementation-level system design, such as how to regulate cross-border transfers of neural data or how to guarantee procedures for obtaining patient consent and requesting data deletion.
The NMPA's approval announcement is similar. The core of the review process, including whether trial data can be independently verified and the scope of adverse event disclosure, is not described anywhere in the published materials. With Neuracle Medical Technology reportedly preparing a listing on the Shanghai Stock Exchange's STAR Market, the faster a company rushes to expand internationally to raise funds, the easier it becomes for the question of where patients' neural data will be accumulated, and under what management framework, to be left behind.
A "global coalition" idea that retraces the limits of the Bletchley Declaration
To fill these gaps, experts are calling for a framework of international coordination. Florian Solzbacher, co-founder of Blackrock Neurotech, said in an interview with thedebrief.org, "This is more than a technology race; it is a contest over whose values will lead the world." Martin Schüttler, CTO of CorTec Neuro, also pointed out in the same interview that China is becoming positioned to lead in setting what is effectively the "gold standard." Kerry Bowman, a bioethicist at the University of Toronto, warns of the risk that "regulatory arbitrage" will arise in regions with thin regulation.
But this kind of "global coalition" idea has a close precedent. In AI, 28 countries and the EU signed the "Bletchley Declaration" at Bletchley Park in the UK on November 1 and 2, 2023. Building on it, the Seoul AI Summit in May 2024 agreed to set up an international network of AI safety institutes, and its first meeting was held in November of that year. The declaration itself is not legally binding, and the only thing guaranteeing effectiveness is voluntary cooperation among countries. If the BCI "global coalition" proposal follows the same path, it may inherit even the Bletchley Declaration's weakness of limited effectiveness.
Moreover, the "global coalition" proposal itself does not yet have even the specificity of the Bletchley Declaration. The Bletchley Declaration fixed its signatories at 28 countries and the EU, and half a year later led to the next step, agreement at the Seoul Summit to establish a network of safety institutes. The BCI "global coalition" idea, by contrast, has yet to address institutional design questions such as who would take part, what measures would be taken in case of violations, and which existing international body would serve as the secretariat. Before the content of a framework, only the ideal is running ahead.
Charles Lieber, a former Harvard University professor who was convicted in 2021 of concealing his ties to a Chinese talent recruitment program, entered China on April 28, 2025, and as of May 1 became a researcher at the Shenzhen Medical Academy of Research and Translation (SMART) and founding director of "i-BRAIN," a Chinese state-funded brain research institute. i-BRAIN is said to even be equipped with deep ultraviolet lithography equipment made by ASML. When even a person with a criminal conviction is welcomed into the core of a national project, it is hard to stop the reverse flow of talent and technology with legal sanctions alone. With countries' intentions this divided, making binding international rules effective remains far from easy.
The question facing Japan, which has no precedent for an invasive BCI
It is not that Japan has no BCI- or BMI-related medical device certification at all. The non-invasive "LIFESCAPES medical BMI (finger type)" obtained medical device certification in March 2024, and from June of that year insurance coverage began under a category called the "exercise volume increase device addition." However, this is a non-invasive motor assistance device that reads signals from outside the scalp, and there is not yet a single case in Japan of regulatory approval for an invasive or semi-invasive BCI, such as NEO, that implants electrodes beneath the skull.
Legislation that directly targets neural data also remains untouched within the framework of personal information protection law. While Chile specially protects brain activity itself through a constitutional provision, Japan's legal system has no provision that gives brain-derived data a special status. With no real-world examples of invasive BCIs in Japan yet, the precedents regulators can refer to exist only overseas.
The risk of "regulatory arbitrage" that Bowman warns of grows larger in regions where regulation is thin. There is no guarantee that Japan is exempt, and the further BCI commercialization advances overseas, the more likely it is that imported products and international data flows will get ahead while domestic debate fails to keep up. Given that China is moving with concrete deadlines, core technology breakthroughs by 2027 and the nurturing of global companies by 2030, Japan's grace period is not long either.
If imports of overseas-made BCIs and international joint trials begin moving ahead while the regulatory vacuum persists, patients' neural data could end up being sent via the cloud to analysis infrastructure abroad. When that happens, it will be too late to debate after the fact which jurisdiction the data falls under. Lacking the accumulated discussion that Chile's neurorights legislation and the MIND Act represent, Japan has to build the very foundations of institutional design from scratch.
Even if binding rule-making cannot move quickly, the contest over standards and values has already begun. The US has taken a step in the form of the MIND Act, at the level of an investigation order, but neither the US, China, nor Japan has yet reached the constitutional-level neurorights protection Chile achieved in 2021. Whether that investigation results in a concrete regulatory proposal will be the next turning point. What NEO's approval has put in front of us is not the outcome of a contest. It is the present state of affairs: the initiative in setting standards has not yet settled into anyone's hands.