中国脑机接口商业化竞赛：政策扶持与临床突破

• This is Mr. Zhang. He asked us not to film his face for privacy reasons. A little over eight years ago, an accident left him paralyzed from the neck down, but here, he's controlling his wheelchair with his mind. - He's even got a few tricks up his sleeve. Last year, Mr. Zhang had surgery to get a brain-computer interface, or BCI from a Shanghai-based startup called NeuroXess. - Training is part of a new wave of clinical trials in China, hoping to make the next breakthrough in BCI tech. - US companies certainly had a head start in developing these systems and testing these systems in humans, but China is certainly catching up very rapidly. - The race to get BCI's to market is on and the stakes for one well-known tech CEO are high.

Like saving humanity from the robocalypse high. - Ultimately, I think this helps mitigate the civilizational risk of artificial intelligence. - AI doom scenario aside, there's a chance BCI's could move beyond just medical purposes and further enmesh humans and technology. But first, they've gotta get to market. - Tiny electrodes are carefully cemented to the scalp to detect and record the delicate pulses from the human brain. - There's been a lot of buzz around brain-computer interfaces in the past few years, but the technology behind them isn't all that new. - So a brain-computer interface is a piece of technology that allows your brain to communicate directly with a computer or another sort of piece of electronic equipment bypassing the rest of the body. - The term brain-computer interface was coined in the 1970s by an American scientist who showed that one could take electrical signals measured by an EEG and map it onto patterns and then decode how that corresponds to movement. - It wasn't until sort of the early 2000s that we got into humans using these chips to read the activity in human brains. - Over the past few years, it's really entered mainstream consciousness.

Thanks in no small part to the enthusiasm of tech billionaires, but we are at a point where are moving fast in clinical trials and really working towards bringing a product to commercial use. - BCI's work like this: Everything you do, whether that's speak, eat or move your hand is driven by neurons in your brain. Those neurons send electrical signals to each other to create action. BCI's read and decode those signals to control an external device like a tablet or prosthetic limb.

BCI's come in varying degrees of invasiveness. Some need to be surgically implanted and either stick into the brain or sit right on top of it. Others are placed on top of the scalp. Regardless of how you measure it, brain data is noisy. That's where AI comes in. BCI companies use algorithms to filter out background chatter and accurately identify the brain's instructions. - The vast rich amount of data that they collect can be super helpful in helping scientists understand the human brain better.

And a lot of researchers believe that this kind of data will help AI researchers build totally different kind of artificial intelligence systems. - In the US, it's estimated that BCI companies have raised $2.75 billion collectively. - Chinese and European companies have raised maybe a fraction of that. - Even so, the Chinese sector is growing. Funding for Chinese brain chip startups more than doubled in 2025 compared to the year before.

And in March of 2026, China approved its first invasive BCI for commercial use, a signal that it's betting big on this tech. - My name's Tiger Tao, and actually Tiger is the direct translation of my first name Hu in Chinese. So it has nothing to do with my personality. - Over the past few years, NeuroXess has been experimenting with different approaches to BCI's. - I was hoping that maybe you could bring the Marva - The device uses a strip of electrodes that sits on top of the cortex that can record signals in the brain.

Those electrodes are connected to a brain chip that does all the signal processing that's attached to an internal control unit that processes and wirelessly transmits data to a computer. It also houses the battery - More than 50 patients have tested different versions of NeuroXess's technology, but Mr. Zhang has lived with it the longest since October of 2025. He spends several hours a day training to learn new skills with his BCI. - It's also interesting that NeuroXess has a collaboration with Xiaomi, one of the leading home appliances companies in China to integrate its BCI with a whole universe of smart home devices. - Other startups have similar partnerships with tech companies so that patients can start to control their devices with their minds.

Mr. Zhang is also training to control a robotic exoskeleton on his hand. According to Tao, this sort of tech could one day help Mr. Zhang walk again. Mr. Zhang will have this implant for an entire year, which is critical to test its long-term safety and effectiveness. After that, Mr. Zhang could remove or keep the BCI, which NeuroXess claims could last three years. - In a sense, it can be a ticking time bomb when something is put in the human brain, which is very sensitive area of the body. - That's because over time, the brain forms scar tissue around the implant, which can damage brain cells and block signals to the device.

NeuroXess opted for a less invasive strip of electrodes to help reduce scarring and maintain a good signal. One challenge? The bar for regulatory approval is very high. - This is gonna be classified as class III medical device, which is the hardest to get in terms of approval. - But they've got a helping hand. Every five years, China's top officials release an outline for key areas of economic and social development.

That support can catapult nascent industries to new heights. For example, China's highlighted new energy vehicles for investment over the past few decades, and today they dominate most of the global EV market. - In the 15th five year plan, BCI was designated as one of six strategic industries of the future. - That means clearer guidelines, faster clinical trial reviews and better coordination. Oh, and a $165 million brain science fund to back the fledgling industry. - So while venture funding for these BCI companies are relatively limited compared to their US peers, this kind of policy support really helps them accelerate their development of BCI technology. - The biggest test will be whether BCI's can actually be a viable business and whether it can live up to all its hype, all its promises when it's rolled out in a commercial, real world setting. - There is a big open question of how these are going to be paid for, what sort of insurance coverage will look like, whether that will be sufficient to to cover the costs. - BCI's are unsurprisingly expensive.

In July 2025, Bloomberg reported that Neuralink estimates each surgery will bring in $50,000. NeuroXess hasn't announced an estimate, but they've already broken ground on a manufacturing facility the company says will make 10,000 devices a year. - I think then, you know, companies and investors are gonna have to see how, you know, large, they really need this market to be. Will it be sufficient to only use this as a medical device?

Or will investors want more of return and want this technology to be available or marketed more broadly? - Mass adoption of this tech, whether invasive or non-invasive, raises questions about the future of neural privacy. - As this technology becomes more widely available, which you know, many people think it will, there are sort of questions to be raised about who has access to this data and what happens to it. - In response to the rapid acceleration of BCI tech, some jurisdictions are passing laws to protect neural privacy. - There are fears that ultimately BCI will be used or abused to alter people's consciousness against their wills or even be deployed in a military setting to create soldiers with enhanced cognitive abilities. - Those nightmare scenarios have some basis in

原文超出正文长度上限，此处截断——上游还有内容，完整版见上方「原文 ↗」。