Thousand-channel probe inserted into the brain, the brain-computer interface industry moves out of the laboratory

Editor’s Note: In the new wave of global technological innovation, future industries are strategic pillars for gaining development initiative. The “14th Five-Year Plan” outline proposes building a comprehensive future industry cultivation system along the entire supply chain, promoting quantum technology, biomanufacturing, hydrogen energy and nuclear fusion, brain-computer interfaces, embodied intelligence, sixth-generation mobile communication, and other emerging economic growth points. Currently, these six major industries are at a critical window from “laboratory breakthroughs” to “industrial explosion.” Starting today, our series “Planning for the Future: New Journey of Industry” will delve into frontline reports, analyze pain points, build consensus, and comprehensively showcase industry dynamics. Stay tuned.

Intern Reporter Yin Jingfei

In the movie “The Matrix,” the protagonist only needs to insert a connection cable behind the head to instantly “download” combat skills and enter a virtual world. This sci-fi scene is approaching reality. However, today’s connection cable has become a slender electrode probe, not for fighting machines, but for reading the brain and curing diseases.

American entrepreneur Elon Musk’s Neuralink has announced the start of large-scale production of brain-computer interface devices, planning to achieve near-fully automated implantation surgeries by 2026. Across the ocean, China’s brain-computer interface industry is also accelerating: policies are intensively supporting, clinical trials are speeding up, and capital is rushing in. From Peking University’s incubation of the Qian Tongdao neural probe to breakthroughs with ultrasonic, flexible electrode, and other diverse technical routes, a competition around “dialogue between the brain and machines” is unfolding.

Industry Segments:

The Revolution of a Thousand Channels on a Single Needle

Drilling a hole in the crown of the head and implanting a device the size of a coin—this is the reality being advanced. The essence of brain-computer interfaces is establishing a direct communication channel between the brain and external devices. Imagine the brain as a vast internet composed of electrical signals, neurons as individual websites, and electrodes as frontline interpreters.

Based on the depth of electrode implantation, brain-computer interfaces are classified into three types: non-invasive, which uses scalp-worn devices to record and interpret EEG signals; semi-invasive, which involves implanting outside the skull or cortex without touching neurons; and invasive, which involves craniotomy to implant electrodes inside brain tissue.

Deep Brain Medical, incubated by Peking University’s Yangtze River Delta Future Technology and Life Health Research Institute, initially developed a neural probe with 1024 signal acquisition channels. This probe not only appeared in the journal Nature Neuroscience but also achieved the world’s first recording of full-depth, high-throughput neuronal activity in a macaque brain.

Liu Xiaojun, head of the brain-computer interface project at the institute, told Securities Times that the probe can be likened to an “earpiece” inserted into the brain, responsible for capturing neural signals; the “channels” are like “radio receivers” (similar to microphones) on the earpiece, with more channels capturing richer signals. One of Neuralink’s core advantages is its R1 surgical robot, which implants 64 slender probes, each equipped with 16 “microphones,” totaling 1024 channels, requiring 64 insertions into the brain. Deep Brain Medical’s approach integrates 1024 “microphones” into a single probe, achieving the same 1024 channels with much less brain damage and easier implantation.

This breakthrough is thanks to its unique “winding” process. Liu explained that traditional flexible electrodes are limited by manufacturing techniques, making it difficult to increase channel count. Deep Brain’s team first fabricates a two-dimensional flexible film densely packed with electrode sites, then uses precise micro-manipulation to wind it into a slender probe—like rolling a flat sheet into a Swiss roll—ensuring the tip has enough recording points while internal space allows wires to “go their own way.”

While the “Swiss roll” probe opens the door for technological innovation, China’s brain-computer interface industry has moved from scattered efforts to full-scale deployment. Before 2020, non-invasive methods dominated; from 2021 to 2024, according to Zhiyao Bureau data, 18 core brain-computer interface companies were established, with invasive methods rapidly emerging; after 2025, the route is diversifying. In January 2026, Gestalt was founded, focusing on ultrasonic non-invasive reading and writing of brain signals, filling a domestic technological gap. Han Xinyong, general manager of Nanjing Zhongke Yixiang Technology, told Securities Times that relying on the company’s new intention decoding paradigm, only about 10 channels are needed to decode high-precision 3D movement trajectories, breaking the industry’s “channel worship.”

According to Zhiyao Bureau, as of February 2026, there are 47 core brain-computer interface companies in China. Data from Qichacha shows that as of January 2026, there are 654 related enterprises nationwide. Industry development is concentrated geographically and by main players, with Beijing, Shanghai, and Hangzhou forming a “tripod,” and Shenzhen following closely. Scientific research powerhouses such as Peking University, Tsinghua University, and the Chinese Academy of Sciences are the sources of innovation, with private capital becoming an important pillar. Chen Tianqiao’s Tianqiao Brain Science Institute and Neuralink started simultaneously; Tianqiao’s brain虎科技 is the only domestic company capable of decoding both movement and Chinese language signals. In 2026, he also co-founded Gestalt, focusing on ultrasonic non-invasive routes.

Commercialization Challenges:

Clinical, Supply Chain, and Multiple Bottlenecks

Currently, the domestic brain-computer interface sector is experiencing policy support and accelerating clinical trials. So far, over 40 hospitals nationwide are conducting clinical experiments. Research reports show that non-invasive methods, due to lower technical barriers and faster commercialization, account for 78% of the market share. However, industry consensus holds that invasive methods are the ultimate future—especially in medical applications.

But in the invasive track, commercialization difficulties are increasingly evident.

First, clinical and certification hurdles.

The development of brain-computer interfaces follows the progression of “core technology—product definition—application scenarios,” with clinical trials being a “must-win” for all companies. Bo Ruikang, which is rushing to list on the STAR Market, is deeply tied to the medical device registration cycle. In March 13, this year, the company’s implantable brain-computer interface system for motor function compensation was approved by the National Medical Products Administration, becoming the world’s first invasive brain-computer interface medical device to enter clinical use. Brain虎’s “super factory” project broke ground in January this year, with plans to obtain certification within 2-3 years.

A senior employee of a leading invasive brain-computer interface startup told Securities Times that most non-invasive devices are Class II medical devices, while semi-invasive and invasive devices are Class III. Few domestic companies can obtain Class III medical device certification, which industry insiders say takes 3-5 years. Most projects are still in IIT (“exploring frontiers”) stage, with some in GCP (registered clinical trial) stage, but the costs of clinical trials are kept secret.

Class III medical devices are “high risk,” requiring strict regulation—long approval cycles, high standards, and almost zero tolerance for errors. A director of investment at a listed medical company told Securities Times that clinical trials for such devices require meeting regulatory sample size requirements, costing at least hundreds of millions of yuan. Certification is time-consuming and expensive, forcing many startups to give up clinical trials. Peng Lei, founder of Gestalt, told Securities Times that for new categories like ultrasonic brain-computer interfaces, global regulatory pathways are still unclear, and companies face the risk of extended commercial return cycles.

Second, industry chain pressure.

A medical device registration certificate is just the surface hurdle. Yang Yang, deputy director of Tianqiao Brain Science Institute (China), told Securities Times that transforming lab technology into stable, mass-producible medical devices is the most direct “bottleneck”—from biocompatibility of flexible electrodes, heat dissipation of implants, to surgical robot support, any failure in detail means system failure.

The brain-computer interface industry chain is long and complex. Upstream includes flexible electrode materials, low-power EEG chips, biocompatible packaging; midstream involves system integration, neural encoding-decoding algorithms; downstream connects to medical device registration, insurance reimbursement, and rehabilitation services.

Li Chun (pseudonym), deputy director of a Jiangsu-based brain-computer interface research institute, told Securities Times that electrode technology varies in difficulty. Although supported by mature supply chains, few companies have truly breakthrough capabilities. Chips are seen as a potential “bottleneck”—as electrode density increases, high-bandwidth signal conversion becomes a necessity. The midstream technical routes are diverse, but the signal decoding and algorithm segments are overly hyped—many companies claim to combine AI decoding, but very few can accurately implement it; most remain at “performance” clinical demonstration stage.

Third, clinical applications have yet to form a complete commercial cycle.

“Within three to five years, invasive brain-computer interfaces will not form a mass consumer market; their commercial value will be limited to urgent medical scenarios like paralysis and Parkinson’s disease,” Liu Xiaojun said.

“Many investors without medical backgrounds misunderstand the market size, thinking it’s huge, but when broken down by specific conditions, the market size is mostly tens to hundreds of billions, not trillions or tens of trillions,” the aforementioned medical company investment director told Securities Times. Clinical diagnosis and treatment are categorized by disease cause; for example, in post-stroke paralysis, it’s necessary to distinguish between stroke-related paralysis and spinal cord injury paralysis. Without such segmentation, clinical applications may become disconnected from hospital diagnosis and treatment processes.

Invasive methods target paralysis, Parkinson’s, epilepsy, but these diseases already have mature therapies, and brain-computer interfaces are not the only option. “Invasive brain-computer interfaces are essentially complex neurosurgical procedures, affected by doctor resources and recovery speed, so their popularization will not be rapid; initial costs are high, and patient acceptance needs time to cultivate,” the investment director added.

Deeper constraints include lagging basic science and ethical red lines. Li Chun told Securities Times that humanity’s decoding of brain neural signals is still in the “blind men feeling the elephant” stage. More importantly, as technology advances from “reading” to “writing” signals and intervening in the brain, commercialization will touch the most sensitive ethical boundaries. The boundary of “consciousness privacy” is no longer just a sci-fi topic; its “double-edged sword” effect could trigger public panic and tighter regulation.

Funding Battles:

Polarization and Lack of Patient Capital

As brain-computer companies enter clinical trials, capital demand becomes more urgent. Research reports show that in 2025, there were 24 financing events in the brain-computer interface field, totaling over 5 billion yuan, mainly focusing on invasive routes, with 60% of funding in Series A and pre-Series A rounds.

While top industry players advance rapidly, many companies struggle with funding. “Polarization, the primary market is very dry or very wet,” a healthcare investor told Securities Times.

Reviewing recent capital trends, Liu Xiaojun divides it into two phases. The first wave started in 2016-2017 with Neuralink’s founding, when domestic capital mainly focused on non-invasive methods. Around 2021, due to slower-than-expected commercialization, enthusiasm cooled. The second wave began in late 2024, mainly driven by policy support. Early 2025, a ladder medical company raised 350 million yuan, exemplifying this trend.

Peng Lei revealed that the funding needs of brain-computer companies far exceed those of typical startups. Neuralink has raised about $1 billion, and domestic companies also require hundreds of millions to over a billion yuan.

Such enormous funding needs highlight the real gap between Chinese and foreign companies—core lies in absolute R&D investment. Liu Xiaojun is aware: “When cutting-edge technology moves toward commercialization, a single technical advantage is not enough to form an absolute barrier; sustained R&D investment is key.” Using Neuralink as an example, “verifying a critical detail of a key technology may require a dedicated team. When you study it, you find they do everything well.” In contrast, many domestic companies in recent years raised only a few million yuan, with teams of only dozens.

Who is willing to pay for such “cost-no-object” investment? Brain-computer interfaces are high-risk, long-cycle advanced technologies, demanding high patience from capital. Most domestic capital still prefers short-term gains. Chen Tianqiao, founder of Shanda Group and Tianqiao Brain Science Institute, as a representative of patient capital, clearly states support for Brain虎科技 for “30, 50 years, even without commercial returns,” but such cases are rare domestically.

Will this wave of enthusiasm cool again? Liu Xiaojun believes the industry has entered a “big wave of sand washing” rational stage. “Capital enthusiasm may not completely fade but will become very rational.” He predicts that in the next 5-7 years, some brain-computer companies will be eliminated due to “inability to raise funds.” Brain-computer interface is a fundamentally long-term industry; no company can turn a profit in 3-5 years, meaning all will rely heavily on fundraising in the coming years. “Storytelling companies will be淘汰, and those lacking core technology or financing ability will also fall.”

A founder of an invasive brain-computer interface company supported by research institutes admitted, “Lack of funds is the core contradiction. Fundraising must ensure smooth development and keep most equity in the hands of the initial team, sometimes requiring compromises. Facing the upcoming industry reshuffle, we want to be among those who can survive.”

Fortunately, now founders of Tencent, Alibaba, and others are investing personal funds into basic research, injecting more patience capital into the industry.