Recently, XS Vision, an invasive brain-computer interface (BCI) company, announced the completion of its seed funding round, raising nearly RMB 100 million. The round was led by Matrix Partners China, with participation from Starlink Capital, Yanyuan Venture Capital, and Shuimu Ventures. This combination of “top-tier venture capital firms + industry/academic capital” represents a relatively high-caliber starting lineup for a deep-tech startup at the seed stage.
It is reported that the funds raised will be used to develop invasive visual cortical prostheses (VCPs), with a strong push toward advancing visual reconstruction algorithms, developing integrated fully implantable brain-computer interface (BCI) systems, conducting preclinical validation, and expanding the core engineering team. The lead investor, Matrix Partners China, was established in 2008 and is a Chinese venture capital firm managing over RMB 70 billion. Recently, the firm has been actively investing in the BCI sector. In addition to backing XS Vision (which develops visual BCI chips with ten-thousand-channel capacity), it has also invested in other innovative companies such as Huachao Shenkong (which specializes in AI-driven ultrasound-based non-invasive BCIs), clearly demonstrating a preference for non-invasive neuromodulation and high-throughput chip technologies.
XS Vision, established in February 2026, is an innovative enterprise dedicated to the foundational hard technology of invasive brain-computer interfaces (BCIs) and their clinical translation. The company specializes in a technical route combining high-throughput BCI chips with neural encoding-decoding algorithms, which determine the upper limits of brain-computer interaction. Spun out from an interdisciplinary BCI research team at Peking University and ShanghaiTech University, its founding members have deep expertise in visual neural mechanisms, language decoding, BCI chips, and embodied intelligence.
Among them, founder Wang Qian is an Associate Researcher at the School of Psychological and Cognitive Sciences, Peking University, and Co-Principal Investigator (Co-PI) at the IDG McGovern Institute for Brain Research. Co-founder Li Yuanning is currently a Professor at ShanghaiTech University; he holds a Ph.D. in Artificial Intelligence and Neuroscience from Carnegie Mellon University and conducted postdoctoral research at the UCSF Edward Chang Laboratory, a top-tier laboratory for brain-computer interfaces. Another co-founder, Zhu Yixin, hails from the Institute for Artificial Intelligence at Peking University, specializing in cognitive reasoning and embodied intelligence, thereby providing algorithmic support for the company’s long-term vision of “thought-based interaction + smart devices.” Furthermore, the team includes several young principal investigators from the China Brain Project as well as hardware chip experts with industrial-grade tape-out experience.
| The product is not yet finalized, but nearly 100 million yuan in funding has already poured in. What has attracted the capital?
Unlike the motor function reconstruction (centered on “reading” brain commands) that has garnered the most attention in brain-computer interfaces over the past few years, visual reconstruction, the area entered by XS Vision, faces a fundamental task: re-encoding external visual information into stimulus signals interpretable by the brain, directly “writing” them into the visual cortex.
The underlying principle involves capturing environmental information via a camera, where algorithms extract key visual features such as contours and position. These features are converted into specific electrical pulse patterns that stimulate the visual cortex through implanted electrodes, inducing "phosphenes" in the patient. If the technology matures, enabling control over the arrangement and combination of light points, patients may be able to perceive lines, letters, and even simple shapes.
The challenge of this approach lies in the fact that the brain is not a pixel-level display; current spread, site interference, and fluctuations in the brain’s intrinsic state all affect stimulation efficacy. Therefore, the system requires an encoding-decoding algorithm capable of real-time “reading” of brain states and adaptive adjustment, thereby establishing a closed-loop “read-write” system.
In parallel, the company’s other approach focuses on linguistic intent decoding for patients with aphasia or severe motor impairments. By reading neural activity in brain regions associated with language and articulation, this method converts intent into text or synthesized speech; for Chinese, additional processing of tones, syllables, and context is required. These two approaches—one for “writing” and the other for “reading”—target different indications yet rely on similar underlying capabilities: high-throughput neural signal processing, precise encoding and decoding, specialized implantable chips, and human intracranial electrophysiological data.
Underpinning these foundational capabilities are three highly scarce technological accumulations of XS Vision.
First, the barrier of human brain visual data. The team has continuously accumulated invasive electrophysiological and electrical stimulation data from the human visual cortex in over 100 cases for more than ten years. It is one of the few teams internationally that possess single-neuron recording data from thousands of human visual cortex neurons, forming an irreplaceable core foundation for visual reconstruction.
Second, decoding algorithms at the international state-of-the-art (SOTA) level. Core team members have long been engaged in cutting-edge research at top-tier international brain-computer interface laboratories, achieving internationally leading algorithmic performance, thereby providing a foundational guarantee for engineering and product commercialization.
Third, an independently controlled 28nm advanced-process chip and a fully implantable system. Against the backdrop of domestic brain-computer interface (BCI) chips generally adopting mature process nodes, the company has achieved systematic breakthroughs in this direction—XS Vision is one of the very few teams in China to have completed tape-out of a 28nm-class BCI chip (benchmarking against Neuralink’s 22nm S2 chip). This solution addresses four core requirements: compact size, low power consumption, high throughput, and bidirectional interaction. The company has already completed the tape-out of a 256-channel chip, with the long-term goal of developing a bidirectional BCI system featuring ten thousand channels and an integrated implantable design. The company defines this solution as a “neural graphics card,” which refers not merely to a single chip, but to a comprehensive platform integrating implantable chips, high-density electrode arrays, decoding algorithms, wireless transmission, and real-time feedback calibration. According to the roadmap, the assembly of the first-generation prototype is expected to be completed by the end of 2026.
Furthermore, building on the aforementioned achievements, the company has outlined a clear pathway for clinical translation. In the near term, it will complete preclinical validation of the fully implanted device’s safety, biocompatibility, and electrical reliability. In the medium term, it will initiate the first-in-human clinical studies, prioritizing patients with acquired blindness who retain relatively intact visual pathways. In the long term, it aims to expand into broader indications for neurological disorders, gradually establishing a scalable cortical visual prosthesis technology platform. Thus, what attracts investors is likely this comprehensive narrative spanning from data to algorithms, from chips to systems, and from short-term validation to long-term expansion.
