Mindtrix Secures $700M Cumulative Funding: Can Brain-Computer Interfaces 'Write' Chinese Characters?

Embodied Emergence has learned that Mindtrix, a domestic brain-computer interface (BCI) company specializing in visual reconstruction, recently completed its Pre-A series financing round, raising a total of RMB 350 million. Combined with its previous angel round financing of RMB 150 million, the company’s cumulative fundraising has reached RMB 500 million.

This latest round of increased capital investment has propelled this company, which is deeply committed to the visual cortex implantation pathway, into the top tier of global funding in the field of visual restoration. It has also once again brought “writing vision into the brain,” a technology still in the clinical research stage, into the industrial spotlight.

As substantial capital flows in, the company has achieved a structured perceptual reproduction of the Chinese character “八” in human investigator-initiated trials (IITs), a breakthrough that has reignited industry discussions on invasive brain-computer interfaces in China.

01

Why Dare to Tackle the Most Challenging Brain-Computer Interface Track?

Mindtrix was founded in October 2024. Its founder, Dr. Liu Bing, holds a Ph.D. in neurobiology and brings over ten years of experience in non-human primate brain-computer interface (BCI) research. He completed postdoctoral fellowships at several leading overseas laboratories before returning to China to establish an interdisciplinary team with expertise spanning neuroscience, electronic engineering, and bioengineering. The company operates under a dual-hub model, with industrialization based in Suzhou and R&D centered in Beijing, focusing on the technically challenging pathway of visual cortex-based visual reconstruction.

From the perspective of embodied emergence, a team’s strategic choices inherently involve industrial trade-offs. While most brain-computer interface (BCI) startups in the market prefer to prioritize directions with lower technological implementation risks, Mindtrix targets individuals with complete blindness caused by eyeball destruction or severe optic nerve damage—conditions that retinal prostheses cannot address. This patient population lacks mature medical solutions, representing an unmet need in healthcare.

With two consecutive rounds of increased investment, investors are not only impressed by the technical demonstration results but also value the team’s long-accumulated expertise in neural experimentation. Given the lengthy clinical translation cycle and substantial capital requirements, the company must rely on continuous capital infusion to advance validation efforts.

02

What Breakthrough Does Chinese Character Recognition Signify?

Many external reports have simplistically characterized the “Identify 7, then Identify 8” task as a mere symbolic demonstration, but the qualitative leap of this IIT trial lies in the iterative refinement of its encoding logic.

The Arabic numeral “7” features simple strokes and limited encoding complexity; in contrast, the Chinese character “八” requires coordinated electrical stimulation across multiple dimensions, including stroke distribution, topological relationships, and visual center of gravity. This signifies that the system has achieved higher-dimensional validation, advancing from simple light spots and single symbols to complex graphic and textual information.

Subjects can perceive structured Chinese characters, provided that a precise mapping between stimulus parameters and the brain’s perceptual outcomes is established, with encoding strategies continuously optimized through closed-loop feedback.

The company publicly disclosed that it has established China’s only database for human visual cortex reconstruction, which is the sole one to have achieved “shape + color + grayscale + Chinese character writing + 100% target hit rate.”

Embodied Emergence: The True Bottleneck in Visual Reconstruction Lies Not in Chip Hardware, but in the "Stimulus-Perception" Paired Data Accumulated from Human Experiments

Data obtained from real human implants of this kind are scarce globally. Constrained by ethical considerations, limited sample sizes, and lengthy acquisition cycles, the rate at which such data accumulate directly determines whether subsequent reconstruction can expand from individual characters to complete words and scene images. While hardware can be procured and iteratively upgraded, there is no shortcut to replicating authentic datasets of human brain perception.

The global visual restoration landscape has diverged into two mainstream technological pathways, with distinct differences observed across commercialization timelines and target patient populations.

Science Corp, a US company pursuing the retinal implant approach, has obtained CE certification for its product in the European Union and achieved small-scale commercialization. However, this solution imposes strict requirements on the preservation status of the patient’s eyeball and optic nerve, resulting in clear limitations in its indications.

US-based Neuralink’s Blindsight, which pivots to direct stimulation of the visual cortex, has received FDA Breakthrough Device designation. Human clinical trials are expected to commence in 2026. This approach bypasses damaged eyeballs and optic nerves, potentially benefiting individuals with total blindness due to optic nerve damage; however, intracranial implantation also entails greater clinical safety challenges.

Domestic companies are also advancing R&D along separate pathways. Nuoxinjia is simultaneously developing two technological lines targeting the retina and the visual cortex, with clinical preparations underway for both pipelines. Mindtrix, on the other hand, is fully focused on visual cortex implantation and is the first team in this domestic sector to complete multidimensional investigator-initiated trial (IIT) validations in humans, including the perception of shapes, colors, and Chinese characters.

Retinal and visual cortex approaches are not mutually exclusive alternatives; they target blind patients with different types of injury and will coexist and develop in parallel over the long term.

03

How Far Is It from Widespread Public Adoption?

According to the company’s public roadmap, it will conduct China’s first investigator-initiated trial (IIT) in completely blind patients in 2026, advance to formal registrational clinical trials in 2028, and strive to launch its first-generation product by 2030, with several more years remaining before true commercialization is achieved.

The RMB 500 million in financing will provide fuel for clinical R&D; however, invasive medical devices entail substantial resource consumption due to clinical trial costs, ethical review processes, and long-term implant safety assessments, with funding being only one component of the equation.

Embodied Emergence posits that the current industry is prone to a common misconception, habitually judging the maturity of brain-computer interface (BCI) products by funding scale and flashy demonstrations.

Visual reconstruction is not a rushed product cobbled together with capital; database iteration must be slowly refined through case-by-case human trials. While the capital market may anticipate short-term milestones, for invasive medical brain-computer interfaces, demonstration efficacy cannot be equated with product usability.

The industrial value of this technology extends beyond merely creating a “bionic eye.” This closed-loop database of visual cortex encoding will also feed back into fundamental neuroscience research, helping humanity further understand how the brain’s visual cortex processes external signals.

Of course, we must also view the boundaries rationally. The first-generation products will most likely only be capable of contour recognition and obstacle avoidance; there remains a significant gap before they can reconstruct high-definition, full-color real-world visuals. Capital has provided the fuel for the industry to accelerate, but the brain itself remains the most challenging subject to conquer in research.

       Original Title: “Mindtrix” Has Raised a Cumulative Total of 500 Million Yuan! Can Brain-Computer Interfaces Really “Write” Chinese Characters?