Brain-Computer Interface: The Future Is Here? Challenges and Prospects of China's 'BCI Four Dragons'

As 2025 is nearing its end, brain-computer interface (BCI) technology is gradually becoming a new technological hotspot and focal point. Recently, the November issue of *Pujiang Technology Review*, a well-known Chinese tech commentary journal, launched a special series on brain-computer interfaces by renowned BCI expert Academician Zheng Hairong. The article mentions that we are standing at the starting point of a new era. The widespread adoption of the internet and social media has turned every individual into a creator of data, while the rise of artificial intelligence technology has transformed this vast amount of data into unprecedented means of production. Against this backdrop, brain-computer interface (BCI) technology is moving to the forefront, becoming a crucial bridge connecting biological intelligence with machine intelligence, and may even redefine the boundaries of human cognition and interaction. The article explores how, after a century of exploration, brain-computer interface technology is now on the verge of transitioning from laboratories to large-scale industrialization, discussing the possibilities and future of both invasive and non-invasive BCI technologies.

In the brain-computer interface field, every company faces a "trilemma": how to balance signal quality, invasiveness risk, and market accessibility? Different choices determine the technical path, business model, and even ultimate success or failure.

The article selects four leading brain-computer interface companies in China, including both invasive and non-invasive types: NeuroXess, StairMed, JiXieJi, and BrainCo, as the "Four小龙 of Chinese Brain-Computer Interface" for detailed investigation and analysis. It thoroughly reviews their respective development histories and strategic routes. By examining the technical and commercial strategies of these four leading companies, the article further discusses the potential for industrial implementation and future development of the entire brain-computer interface industry. Let us take a look at whether the "Four小龙 of Chinese Brain-Computer Interface" can successfully represent the Chinese brain-computer industry in achieving industrial success and paving the way for a unique future.

NeuroXess

NeuroXess, a well-known representative of Shanghai-based brain-computer interface companies, has completed cumulative financing amounting to hundreds of millions of yuan. Its main technology and products focus on invasive brain-computer electrodes. Tao Hu, the founder, transitioned from being a scientist to an entrepreneur. His reasoning is: "China doesn't lack scientists who publish papers, but it lacks people who can bring technology to real-world applications." The mismatch and vast gap between scientific research and clinical needs, as well as between top-tier academic papers and patient demands, requires scientists like him to dive into brain-computer interface entrepreneurship to bridge the divide. By the end of 2024, Tao Hu made a bold decision to step down as the deputy director of the Shanghai Institute of Microsystem and Information Technology under the Chinese Academy of Sciences in order to fully commit—"ALL IN"—to NeuroXess.

As he put it, "The development of brain-computer interface (BCI) technology is progressing so rapidly that dedicating myself full-time to NeuroXess is necessary to avoid the dispersion of resources between scientific research and entrepreneurship, thereby accelerating the industrialization of BCI technology." In the cutting-edge scientific research market, the company offers more technically aggressive deep electrodes. These electrodes can be inserted into the internal layers of the cerebral cortex, enabling precise capture of individual neuron firing activities, providing neuroscientists with high-precision tools to explore the mysteries of the brain. In the clinical medical market, NeuroXess adheres to the principle of "patient benefit and safety first." Since any penetration of brain tissue can trigger inflammatory responses and pose long-term safety risks, for patients requiring lifelong implants, this must be the highest priority consideration. Therefore, NeuroXess has opted for minimally invasive micro-electrocorticography (µECoG) arrays—electrodes are placed only on the surface of the cerebral cortex without penetrating deeper tissues, significantly reducing surgical risks. This system employs a 256-channel flexible electrode array with a density of 64 channels per square centimeter, which is 64 times that of traditional electrocorticography (ECoG) electrodes. Through semiconductor microfabrication processes, the team achieved "zonal design": ultra-thin mesh recording zones closely conform to the cortex to ensure high-fidelity signal acquisition, while reinforced lead zones withstand long-term mechanical wear from implantation; custom titanium alloy waterproof enclosures further enhance reliability. This design largely achieves a balance of "high throughput, high resolution, and low invasiveness."

Centered around electrodes, NeuroXess adopts a "full-stack" vertical integration model, covering all core aspects from hardware, software, and algorithms to surgical robots. Tao Hu explained that this approach is both an objective necessity in the early stages of the industry and an active choice to build competitive barriers. The current brain-computer interface industry is still in its infancy, far from having a mature supply chain ecosystem like that of smartphones. Most products on the market have channel counts concentrated between 8 and 64, primarily limited by the number of channels available in purchasable chips, leading to compromises in system-level performance. By independently designing a 256-channel dedicated chip, NeuroXess has achieved deep compatibility with electrode arrays, enabling system-level collaborative optimization across multiple dimensions such as signal acquisition, processing, packaging, and power consumption. Tao Hu pointed out, "Manufacturing tens of thousands of electrode channels is not technically difficult, but without matching chips, packaging technology capable of handling massive connection points, and system designs controlling power consumption and heat dissipation, even more electrodes would be rendered useless." NeuroXess is not pursuing isolated "technological showmanship" but rather optimal system performance.

NeuroXess has planned its own commercialization path: accumulating data and cash flow from the scientific research market, achieving clinical application through the approval of medical devices, and possibly entering the consumer-level market in the long term. Currently, it is focusing on深耕 the medical rehabilitation scenarios. The practical path goes from flexible electrodes to silk protein coating, from full-stack R&D to clinical validation, hoping that all these paths will lead to the same goal: making technology truly serve those patients in need.

StairMed

Similarly rooted in Shanghai, StairMed is a brain-computer interface company that focuses on invasive brain-machine electrodes as its core technology. Founded by Zhao Zhentuo, a researcher from the Center for Excellence in Brain Science and Intelligence Technology at the Chinese Academy of Sciences, the company has completed cumulative financing amounting to hundreds of millions of yuan. It also holds the record for the highest single-round financing amount in 2025. Its core technological advantage lies in an independently owned intellectual property — "ultra-flexible micro-nano electrodes" — which aims to fundamentally change the physical properties of electrodes. Made from biocompatible materials like polyimide, the core design concept is to make their mechanical properties closely resemble the soft brain tissue itself.

Zhao Zhengtuo introduced. On March 25, 2025, in an operating room at Huashan Hospital Affiliated with Fudan University, a coin-sized implant was precisely embedded into the skull of a patient with quadriplegia. This marks China's first invasive brain-computer interface system prospective clinical trial (First-In-Man, or FIM), a milestone achievement that positions StairMed as the second company globally, following Neuralink, to publicly announce reaching this critical clinical stage.

At the same time, StairMed has made advanced arrangements in production and manufacturing. It has built a 2,000-square-meter production plant that meets international GMP standards and plans to complete China's first medical-grade brain-computer interface MEMS production base in 2026. Zhao Zhentuo explained the necessity of building their own production line: "Medical scenarios require implants to work for a long time in environments with body fluid corrosion and immune attacks, which is completely different from the usage environment of consumer electronic chips. This requires extensive modification of materials and processes to form a whole new set of standards. This experience and capability cannot be outsourced and must be built from scratch."

However, for the global technological competition in brain-computer interfaces, the ultimate value of the technology will also be determined by many challenges, such as more rigorous clinical data validation, scaled production, regulatory compliance, and commercial sustainability. In Zhao Zhengtuo's ultimate vision, the final goal of StairMed is to achieve a revolutionary improvement in human-machine interaction efficiency, with the brain-computer interface system ultimately becoming the next-generation computing platform.

Zhao Zhengtuo described StairMed's development strategy as "beginning with the end in mind," combining lofty aspirations with practical steps. Starting with medical scenarios that can generate clear value, they aim to address core challenges for populations with urgent needs. The first area is motor control, targeting patients with motor function impairments such as paralysis or amputation. These individuals can regain connection with the world—and even restore work abilities and personal dignity—by controlling computers, smart prosthetics, and smart home devices through thought alone. The second focus is sensory function restoration, such as artificial hearing. For patients who cannot benefit from cochlear implants due to damaged auditory nerves, direct stimulation of the auditory nuclei in the brainstem can rebuild their sense of hearing, enabling them not only to hear sounds but also perceive musical melodies. The third area is deep brain neuromodulation, targeting diseases like Parkinson’s, epilepsy, and depression. Traditional deep brain stimulation (DBS) systems use open-loop stimulation, while StairMed’s brain-computer system enables closed-loop modulation, which involves real-time reading of brain signals to determine pathological states, followed by precise and intelligent stimulation. This results in better treatment outcomes and fewer side effects.

StairMed currently focuses more on the clinical development pathway, planning a timeline for product launch: initiating a nationwide, multi-center, larger-scale clinical trial in 2026 across China, with an expected enrollment of 30 to 40 subjects to provide critical efficacy and safety data for the registration and market approval of its Class III medical device.

StairMed

Unlike the previous two invasive brain-computer interface companies, MaschineRobot, founded by Lu Shuqiang, a serial entrepreneur in the brain and intelligence industry, in collaboration with experts from Stanford and MIT research teams, has chosen a different development path driven by AI technology for brain-computer interaction, accumulating hundreds of millions in financing. The company is positioned to explore technical solutions to the ultimate philosophical questions of memory backup, consciousness replication, and body replacement. MaschineRobot's vision is straightforward and bold: "True Love and Immortality." The logic is: if humans struggle to find true love, then create a robot that can provide it; if humans cannot escape death, then upload consciousness to the digital world. This quintessentially American tech narrative and development model brings a unique approach and business logic to the brain-computer interface industry.

This company has chosen a dual-market commercial route for both consumer and enterprise-level products. In the consumer market, NeuroXess is targeting broader non-clinical scenarios such as health, entertainment, art, and education. Under the brand "BCI-X," it has launched a series of neurotechnology entertainment interaction products. At the same time, it has adopted parallel technology routes for both invasive and non-invasive approaches. The core of its invasive projects is the "S-BCI" chip platform, which focuses on developing multiple high-throughput invasive chips for use in scientific research and clinical settings, such as helping ALS patients regain their ability to communicate.

For instance, in 2024, Ji Machinery did not release a robot capable of "romantic relationships," nor did it showcase a technological breakthrough in "consciousness uploading." Instead, its launched products included Dreamgear (Dream Recording and Generation), BCI-Music (Emotion-Generated Music), X-BCI (Enterprise-Level Brain-Computer Hardware Platform), Mind++ (Enterprise-Level Brain-Computer AI Computing Platform), Brain++ (Enterprise-Level Brain-Inspired Computing Platform), and elderly care companion robots. These products appear more like entertainment and health service tools rather than bridges to "immortality."

In 2025, the company will launch its open product solutions, including BCI-love for brain-computer emotional interaction, BCI-sleep for brain-computer sleep interaction, and BCI-control for brain-computer control systems. Additionally, this company has collaborated with Peking Union Medical College Hospital to provide ALS patients with a communication, interaction, and sensory enhancement system based on brain-computer interfaces. The corresponding products are BCI-communication (a brain-computer typing communication system) and BCI-Feeling (a brain-computer sensory enhancement system).

JiXie Machine Targets the B-end Market, Transforming Years of Technical Expertise and Data Accumulation into Platform Development. As a provider of fundamental brain-computer interface infrastructure, it has launched the "X-BCI" brain-computer hardware platform. Positioned as a “brain-cloud” service model, various industry vertical service providers can develop applications on the platform. As an infrastructure for industrial applications, the X-BCI platform consists of brain-computer hardware units, a brain-computer data service platform, and a brain-computer AI computing system. It offers diversified brain-computer hardware with 1 to 64 channels and integrates backend data storage functionality. By leveraging dedicated AI computing models, it enables intelligent processing of brain-computer interaction tasks, allowing users to easily perform one-click invocation and deployment. JiXie Machine positions itself as not needing to become an expert in every vertical field but instead collaborates with leading enterprises across industries through X-BCI. In fields such as film, music, finance, and industrial safety, JiXie Machine provides the technical foundation, while these partners combine their industry knowledge and channel resources to develop brain-computer applications tailored to specific scenarios. JiXie Machine then earns a share of the subsequent profits from these applications, employing this classic "selling shovels" strategy.

Its development positioning initially focuses on consumer-grade and B-end industry service brain-computer interfaces, accumulating data, technology, and capital, then gradually tackling the long-term goals of invasive systems and consciousness digitization. The immediate focus is on non-invasive commercialization while treating invasive systems as a long-term R&D project reserve. JiMech’s artificial intelligence gene sets it apart from other brain-computer teams that enter the field via hardware electrodes or neuroscience. From its inception, JiMech has had a strong AI and data science background, more reasonably positioning itself as an “AI-driven brain-computer interface” with the core methodology being “data + algorithms.” The underlying logic is this: the essence of a brain-computer interface is decoding. EEG signals recorded by non-invasive EEG are extremely complex, weak, and full of noise—non-linear, non-stationary, and with extremely low signal-to-noise ratios—meaning accurately identifying specific intent signal patterns from mixed background noise presents significant challenges. Traditional signal processing methods and simple machine learning models often fall short when handling high-dimensional, high-noise data, which is where modern AI technologies like deep learning play a critical role.

But in the eyes of its founding team, advanced algorithms alone cannot constitute a lasting moat. "In today's increasingly widespread use of AI technology, innovations in model architecture are quickly replicated. The true barrier for StairMed lies in the high-quality, large-scale proprietary datasets that drive these algorithms"—this is what he referred to as the "most expensive" part since the company’s establishment. The collection of EEG data is extremely time-consuming and labor-intensive. Unlike computer vision or natural language processing, which can scrape massive amounts of data from the internet, every valid EEG data point must be obtained under strictly controlled experimental conditions through carefully designed experiments. StairMed has invested enormous resources into this: setting up dedicated data collection labs at universities like Zhejiang University, hiring neuroscientists to lead experiments to ensure the scientific rigor and standardization of the data, and then systematically accumulating over several years to build a vast database supporting its various products. The combination of this proprietary dataset with advanced AI algorithms creates a powerful flywheel effect: high-quality hardware and professional collection processes ensure data "purity" → large amounts of quality data train accurate, robust AI decoding models → superior performance models give rise to breakthrough experience products like Dreamgear → products launched into the market attract more users, generating more real-world scenario data → new data is used for iterative optimization of models, further enhancing the experience. This closed loop of "data-model-product-user-data" forms a self-reinforcing positive feedback cycle.

BrainCo

Headquartered in Hangzhou, BrainCo, one of the "Six小龙" in Hangzhou, is one of the earliest companies in China to enter the brain-computer interface field, having cumulatively completed nearly 1 billion yuan in financing. This company, incubated from Harvard's Innovation Lab, has been clear about its technical route since its establishment in 2015 – choosing the non-invasive path. After nearly a decade of deep cultivation, BrainCo has become one of the leading enterprises in the global non-invasive brain-computer interface field.

It has built a complete technology stack from self-developed electrode materials to artificial intelligence algorithms at the underlying technology level. According to publicly available data, BrainCo achieved the mass production of 100,000 units of a single brain-computer interface product in 2022, achieving significant commercialization results. Behind the choice of the non-invasive approach is BrainCo's assessment of technological value — providing assistance to people with disabilities, patients with neurological diseases, and the general public in a way that is lower risk and more accessible. The surgical risks of invasive brain-computer interfaces, safety concerns related to long-term implantation, and complex ethical controversies make it difficult for them to achieve large-scale adoption in the short term. In contrast, non-invasive brain-computer interfaces lower the threshold for use and can be safely applied to a broader population, including patients in rehabilitation, children with special educational needs, and ordinary people pursuing physical and mental well-being. In BrainCo’s philosophy, user-friendliness is key to rapidly expanding market reach.

According to BrainCo's assessment, the surgical risks, long-term implant safety concerns, and complex ethical controversies associated with invasive brain-computer interfaces make it difficult for them to achieve large-scale adoption in the short term. On the other hand, non-invasive brain-computer interfaces lower the usage threshold and can be safely applied to a broader population, including patients in rehabilitation, children with special educational needs, and the general public pursuing physical and mental well-being. In BrainCo's philosophy, user-friendliness is the key to rapid market expansion. A partner of BrainCo stated, "Bringing brain-computer interfaces back to the human dimension is not about chasing flashy concepts but using technology to help more people rebuild confidence and embrace life again—this is where the true power of this technology lies."

BrainCo's choice avoids the current commercialization path of invasive technology, which is full of uncertainties — requiring passing through multiple hurdles such as long-term clinical trials, medical device approvals, and surgical cost controls. This allows BrainCo to bring products to market more quickly, validating technology and accumulating data through real commercial scenarios. Based on signal acquisition and algorithm decoding technology, BrainCo's product lines mainly focus on two core directions: intelligent bionics and intelligent health.

BrainCo's intelligent bionic product series primarily targets disabled individuals, including the intelligent bionic hand and bionic leg. The intelligent bionic hand operates by capturing weak electromyographic (EMG) and neural electrical signals generated on the surface of the user’s residual limb through sensors worn at the end of the residual limb. These signals are decoded in real-time by an embedded AI algorithm to drive the various joints of the bionic hand, enabling highly intuitive mind control. In 2022, BrainCo's intelligent bionic hand received FDA approval for market release in the United States. Additionally, the company has launched a series of intelligent health products aimed at a broader audience, focusing on the intervention, rehabilitation, and general brain health management of brain-related disorders. This includes brain-computer interface training systems for children with autism and attention deficits, as well as wearable devices for sleep assistance, stress relief, and mindfulness training designed for general consumers.

In the future, BrainCo's core direction will focus more on the intervention and rehabilitation of brain diseases, continuing to invest in EEG signal and neural mechanism research, and further exploring the application potential of brain-computer interfaces in areas such as depression and Alzheimer's disease. From a technical architecture perspective, BrainCo has established a complete chain from signal acquisition hardware, AI decoding algorithms, to multi-scenario application systems. From the self-developed solid-state gel electrode materials, to AI algorithms, and then to applications in fields like prosthetics, rehabilitation, and overall health, BrainCo has built a technology system that integrates software and hardware.

As the commercialization process advances, BrainCo is attempting to establish a positive cycle between technology, products, and users: by obtaining user feedback through product usage, optimizing algorithm models, and thereby improving product experience. The BrainCo team believes that the key to whether non-invasive brain-computer interface companies can succeed in future competition lies in the technical feasibility and systematic integration capabilities. They predict that the future development of the brain-computer interface industry will show three trends: The first is integration. Brain-computer interfaces will deeply integrate with artificial intelligence, neuroscience, wearable devices, and other fields; miniaturization of hardware and flexibility of materials will make devices smarter and easier to use. The second is accessibility. Brain-computer interfaces will move from laboratories to broader life scenarios, with increasingly simpler user experiences. The third is standardized development. As brain science research and medical practice deepen, brain-computer interfaces are moving toward systematization and standardization. They will gradually establish clinical usage norms, safety evaluation systems, and ethical standards, making technological development more orderly and reliable.

Through the above analysis, we can see that the four小龙 of the brain-computer interface—BrainCo, NeuroXess, StairMed, and others—differ in their technical positioning, development strategies, commercial routes, and visionary goals. However, they all aim to achieve commercial implementation and breakthroughs in large-scale industrial applications based on their respective development logic.

At the same time, with the Fifteenth Five-Year Plan identifying brain-computer interface as one of the key development industries, regions across China have positioned brain-computer interface as a significant development strategy. Additionally, new outstanding brain-computer startups led by scientists in various directions have also joined this major track based on the industry-university-research transformation route. Examples include Beijing Zhiran Medical, Beijing Heze Technology, Shanghai Borui Kang Medical, Beijing Xinzhi, Shenzhen Weiling Medical, and Mingshi Brain-Computer Technology, each choosing different development paths such as brain-computer electrode materials, brain-computer system platforms, and brain-computer clinical applications for technology and product exploration.

Besides, several well-known brain medical technology companies that have completed mature business model construction, such as Pincer Medical, SceneRay Medical, and Sanbo Brain Hospital, are also actively paying attention to the development of brain-computer interface technology. As the market application and prospects of brain-computer interfaces become increasingly clear, they are expected to directly participate in the product development and commercial launch of the brain-computer interface industry in the future. Meanwhile, some renowned large technology companies such as Huawei, Xiaomi, ByteDance, Alibaba, Tencent, and Meituan are also actively laying out plans for exploring brain-computer interface technology and applications, with the hope of launching related technologies and product application services in the future.

What’s Next for the "Big Four" in Brain-Computer Interface? As more emerging players enter the BCI industry, along with established neuro-medical tech companies and even internet giants joining the race, how will they continue to maintain their technological and commercial edge? And how will they realize their technical practices and achieve commercial breakthroughs? Let’s wait and see!

As Academician Zheng Hairong mentioned, brain-computer interfaces are not a distant concept. From the mind-controlled flying dragons depicted in the early movie *Avatar* to the current reality where paralyzed patients control robotic arms through brain signals to perform delicate movements, this field is continuously achieving significant breakthroughs. Its core mechanism lies in the "decoding" and "encoding" of neural signals in the brain to establish an efficient, bidirectional information exchange channel between the brain and external devices. At present, various technical pathways such as invasive, semi-invasive, and non-invasive approaches have made respective advancements while also facing challenges related to biocompatibility, signal accuracy, and safety. In the healthcare sector, brain-computer interfaces demonstrate profound application potential. They have already been applied in multiple areas including motor function reconstruction, sensory substitution, and neurological disease treatment, bringing new hope to groups such as spinal cord injury patients and the blind. Research teams and enterprises both in China and abroad are actively working to advance brain-computer interfaces from laboratory research to clinical applications and industrialization.

However, the significance of brain-computer interfaces extends far beyond medical rehabilitation. It more profoundly points to an interdisciplinary proposition: how to achieve the deep integration of biological intelligence and artificial intelligence. As *Nature* magazine noted during the commemoration of Turing's 100th birth anniversary, "The fusion and collaboration of the brain and machines is a crucial pathway to achieving artificial intelligence." The future development trajectory will not only include "brain-controlled machines" but may also expand into "brain-to-brain communication," enabling direct information exchange between living organisms and breaking through the physiological and physical limitations of traditional communication. This progress will have profound implications for education, communication, collaborative models, and even the structure of human society.

From the macro perspective of technological evolution, we are currently still in the "data intelligence" phase, where artificial intelligence empowers various industries through big data and large models. In the next phase, technology will move towards "physical intelligence," achieving deep interaction between AI and the physical world. Ultimately, through the bridging role of brain-computer interfaces (BCIs), the "biological intelligence" phase is expected to realize the integration of humans, machines, and the environment, constructing the ultimate vision of intelligent forms. Achieving this vision requires multidisciplinary collaboration and systematic innovation. Fields such as neuroscience, computational science, materials engineering, algorithm design, and ethical governance must be deeply integrated to address technical challenges and build controllable, secure, and reliable BCI systems. Meanwhile, policy support and industrial resource investment are also indispensable, which has become a frontier direction jointly focused on by national scientific and technological strategies and innovation alliances.

In this era full of potential, brain-computer interface (BCI) not only represents a key technology but also triggers deep reflections on the future form of human existence. It reminds us that the true intelligence revolution will ultimately return to humanity itself—expanding cognitive boundaries, deepening self-understanding, and ultimately achieving deeper connections between lives. On the BCI track, there are no shortcuts. Global players have been building barriers through years or even decades of clinical data accumulation. StairMed has chosen "the most difficult path." Whether it can evolve from "a sample in China" to "a global competitor," the data will speak louder than stories in the times to come.

*Disclaimer: This article is submitted by the manufacturer, only for conveying information, and does not represent the viewpoint of 87870.com.

Author:

Freelance Tech Commentator (If there is any infringement, please contact us for removal)）

References:

(1) Pujiang Technology Review, November 2025 Issue

(2) NeuroXess Official Website https://www.neuroxess.com

(3) StairMed Official Website http://www.stairmed.com

(4) Jimei Machinery Official Website https://www.maschinerobot.com

(5) BrainCo Official Website https://www.brainco.cn