From 'Reading the Brain' to 'Writing to the Brain': SiClink Secures Tens of Millions in Funding to Pioneer Bidirectional Brain-Computer Interfaces

Invasive Brain-Computer Interface Startup SiClink (Xi Lian Technology) Recently Completes Tens of Millions in Seed and Angel Financing Rounds, Jointly Invested by BlueRun Ventures China, GL Ventures, and Zhongke Shenguang. Against the backdrop of brain-computer interfaces being included as a key focus area in China’s 15th Five-Year Plan and with many industry insiders viewing 2026 as the inaugural year for accelerated industrialization, this financing is not merely an example of a startup securing capital support; it further reflects a shifting perspective within the capital markets regarding the technological roadmap for next-generation brain-computer interfaces.

In recent years, the focus of the brain-computer interface (BCI) industry has primarily been on “reading the brain”—that is, how to decode motor intentions, linguistic intentions, or other cognitive information from neural activity. Whether it is Neuralink enabling paralyzed patients to control computer cursors with their thoughts, or various speech decoding systems helping patients with amyotrophic lateral sclerosis (ALS) regain communication capabilities, these advancements essentially represent improvements in neural signal readout capabilities. However, as these technologies gradually transition from laboratories to clinical validation, a deeper question has begun to emerge: If BCIs are to become not just medical devices but the next-generation human-computer interaction platform, merely being able to “read the brain” will clearly be insufficient.

The next phase of competition will no longer be solely about the precision of neural signal decoding, but will center on “integrated read-write” capabilities. Those who can establish a long-term stable, high-bandwidth, bidirectional information pathway between silicon-based chips and carbon-based neural tissues will have the opportunity to define the next-generation brain-computer interface platform. SiClink’s chosen entry point—Vision Reconstruction—is precisely one of the most challenging directions in this competition.

Why Is Capital Beginning to Focus on “Bidirectional Brain-Computer Interfaces”?

From the perspective of industrial development patterns, any technology undergoes a progression from “proving feasibility” to “demonstrating utility,” and finally to “establishing scalability.” Over the past decade, the primary objective of the brain-computer interface (BCI) industry has been to demonstrate that neural signals can be effectively read. Today, this capability has been validated in multiple clinical trials. Paralyzed patients have been able to use BCIs to control robotic arms for grasping tasks, input text via thought, and even operate wheelchairs and external devices. This signifies that neural signal decoding has gradually transitioned from a scientific challenge to an engineering problem.

However, for capital investors, merely staying within the domain of motor control is insufficient to support a trillion-dollar industry. Although motor control holds significant clinical value, it primarily serves specific patient populations, resulting in relatively well-defined market boundaries. To enable brain-computer interfaces (BCIs) to expand into broader human-computer interaction scenarios, the fundamental challenge of how information enters the brain must be addressed.

In a sense, reading the brain merely provides the capability of an “input device,” whereas writing information into the brain signifies the acquisition of “display” and “output device” capabilities. Only when a system possesses both read and write functionalities does the brain-computer interface (BCI) truly have the potential to become a new information infrastructure. Therefore, what deserves greater attention behind SiClink’s recent financing is not the amount raised, but rather the capital markets’ endorsement of the “bidirectional BCI” approach.

This also signifies a shift in the logic of technological competition within the industry. The focus is gradually transitioning from breakthroughs in single-point capabilities to the ability to build complete, closed-loop systems. It is shifting from emphasizing signal quality in individual experiments to ensuring stable operation after long-term implantation. Furthermore, the focus is moving from specific functional scenarios to assessing the potential to establish a platform-based technological ecosystem.

Visual Reconstruction: The Most Challenging Frontier in the Field of Brain-Computer Interfaces

Among the various application directions of brain-computer interfaces, visual reconstruction is regarded by many researchers as a challenge at the "Mount Everest" level. The reason lies not in vision being more important than motor function, but in the visual system itself possessing extremely high information density and highly complex neural encoding mechanisms.

Motion control is fundamentally about decoding the user’s “intent to act.” For instance, when a patient forms the intention to move their arm, corresponding neural activity arises in the motor cortex of the brain. Brain-computer interfaces (BCIs) capture these signals, decode them, and translate them into control commands that robotic arms or computers can understand. Although this process is challenging, it essentially remains an issue of information readout.

Visual reconstruction, by contrast, is fundamentally different. It requires the system not only to interpret the information being processed by the brain but also to actively write new visual information into the brain, thereby eliciting corresponding visual perceptions in the user. In other words, the system must function both as a neural signal receiver and as a neural information transmitter.

More importantly, SiClink emphasizes not traditional visual restoration, but visual reconstruction. Traditional visual restoration more closely resembles a unidirectional information transmission model: external cameras capture environmental images, which are then processed by algorithms and used to stimulate the visual pathway, helping patients regain a certain degree of visual perception. The core objective of this model is to restore lost visual capabilities.

Visual reconstruction requires the establishment of a complete closed-loop neural interaction system. The system must not only identify what the user is currently viewing, attending to, and the surrounding environment, but also adjust stimulation strategies in real time based on this information, encoding new visual content into the visual cortex in accordance with neural coding principles. This implies that visual reconstruction is inherently a bidirectional interactive system, rather than a unidirectional information input system.

From a long-term perspective, once visual reconstruction is truly realized, its significance will extend far beyond restoring sight to the blind. Vision is the most critical channel for humans to acquire external information and serves as a key interface for the future integration of the digital and physical worlds. If brain-computer interfaces can achieve stable, high-bandwidth information exchange within the visual system, their underlying technological capabilities are poised to expand into broader domains such as cognitive enhancement, immersive interaction, and neural computing platforms.

Technical Challenges Behind Bidirectional Closed-Loop Systems

Many people have an intuition: Since the industry can already read neural signals and generate perceptions through electrical stimulation, wouldn’t combining these two capabilities achieve a bidirectional brain-computer interface?

In fact, the matter is far more complex than imagined.

Most current systems in the industry are essentially still based on a unidirectional architecture. Recording systems focus on how to acquire high-quality neural signals, while stimulation systems focus on how to generate specific perceptions through electrical stimulation. The problems, optimization objectives, and engineering challenges faced by these two types of systems are not the same.

When a system is required to perform simultaneous read and write tasks, many previously overlooked issues rapidly become amplified. For instance, neural signals drift over time, electrode performance degrades due to long-term implantation, brain tissue mounts an immune response leading to glial scar formation, and stimulation parameters require continuous dynamic adjustment in response to changes in neural states. While these issues might be mitigated through periodic calibration in unidirectional systems, they directly compromise the overall stability of bidirectional closed-loop systems.

Visual reconstruction further exacerbates this challenge. The visual cortex features complex topographic mappings, with distinct regions corresponding to different locations in the visual space. Visual information involves extremely high bandwidth, requiring the system to complete perception, decoding, stimulation, and feedback within an exceedingly short timeframe. Moreover, visual experience is highly dependent on individual differences; identical stimulation parameters may yield vastly different perceptual outcomes across patients.

Therefore, visual reconstruction is not a simple superposition of a “reading system” and a “stimulation system,” but rather a comprehensive technological framework encompassing materials, devices, algorithms, neuroscience, and systems engineering. A weakness in any single component can become a bottleneck that limits the overall performance of the system.

Building Silicon-Carbon Connections at the "Last Micron"

SiClink was established in December 2025. Its name, derived from “Silicon-Carbon Link,” encapsulates the core challenge in the field of brain-computer interfaces: establishing long-term, stable information connectivity between silicon-based electronic systems and carbon-based biological systems.

Dr. He Fei, the company’s founder, currently serves as a Research Professor and Doctoral Supervisor at the Shanghai Institute of Optics and Fine Mechanics, Chinese Academy of Sciences, with a long-standing focus on invasive brain-computer interface (BCI) research. Unlike many research teams that concentrate on algorithmic decoding, Dr. He’s work has consistently centered on the neural interface—a critical component regarded within the industry as the “last micron.”

Over the past decade, He Fei has conducted research on ultra-soft neural interfaces at the University of Texas at Austin and Rice University in the United States, accumulating extensive expertise in highly biocompatible electrode materials, low-threshold neural stimulation, and visual information decoding. Based on publicly published results, his research spans multiple key areas, including neural interface materials, neural stimulation technologies, and visual information processing. This interdisciplinary foundation has become a critical basis for SiClink’s adoption of the visual reconstruction pathway.

Notably, the team has achieved simultaneous recording of visual cortex activity using flexible neural probes in animal experiments, and completed natural scene image reconstruction by combining artificial intelligence algorithms. Meanwhile, the team has also developed an automatic shuttle implantation device and a laser-assisted implantation protocol, aiming to further reduce implantation trauma and improve electrode positioning accuracy.

These efforts may appear fragmented, but they collectively serve a single goal: establishing a long-term, stable bidirectional neural information channel. This is because, regardless of the algorithms or stimulation strategies adopted in the future, if the neural interface cannot operate stably over the long term, all higher-level applications will lack a foundational basis.

From Medical Devices to Information Infrastructure

Looking at the development history of the global brain-computer interface industry, an increasingly obvious trend is emerging: the industry is gradually evolving from "treating diseases" to "enhancing capabilities."

At the current stage, the most viable commercialization path for invasive brain-computer interfaces (BCIs) remains the healthcare market. Whether it involves restoring motor function or speech communication in patients with paralysis, or reconstructing sensory perception for those with visual impairments, these applications have clear clinical needs and regulatory pathways. These scenarios not only validate technical feasibility but also enable companies to establish their initial commercial closed loop.

However, from the perspective of technological evolution, platform technologies that truly change the world rarely remain confined to their initial application scenarios for long. The internet initially served scientific research institutions, smartphones began merely as mobile communication tools, and artificial intelligence was first applied in specific industrial contexts. As these technologies matured, their application boundaries continuously expanded, ultimately evolving into societal infrastructure.

Brain-computer interfaces may also follow a similar path.

The importance of visual reconstruction lies precisely in its simultaneous demands for high-bandwidth neural interfaces, high-density stimulation arrays, long-term implant stability, intelligent encoding and decoding capabilities, and real-time closed-loop control. If these key capabilities can be validated within the visual system, the underlying platform technologies thus established will have the potential to extend to a broader range of application scenarios.

From this perspective, SiClink’s latest funding round represents more than just the growth of a startup; it reflects that the brain-computer interface (BCI) industry is transitioning from “reading neural signals” to “building neural information infrastructure.”

In the next decade, what determines the height of the industry may no longer be who can record more neurons and achieve higher decoding accuracy, but rather who can truly establish a two-way pathway between silicon-based computing and biological intelligence. From one-way reading to bidirectional interaction, from medical devices to information interfaces, from functional compensation to capability enhancement, the next round of competition in the brain-computer interface industry may well begin along this path of "silicon-carbon connection."

The Brain-Computer Interface Community isChina's First Brain-Computer Interface (BCI) Industry Service Platform. Pioneer and Leader in Domestic Brain-Computer Interface New Media。We provide the following services to enterprises, research teams, investment institutions, and industry practitioners:

Publicity and Reporting:Report on corporate updates, technology insights, and product introductions through articles, short videos, and live streams to enhance visibility and industry influence.。

Resource Matching:Match resources such as capital, supply chains, clinical institutions, and channel partners based on needs to facilitate genuine connections and promote collaboration.

Technology Transfer:Assist the technical team in identifying industry partners, investors, and application scenarios to facilitate the translation of technology into products.

Event Planning and Execution:Undertake the planning and execution of online and offline roadshows, salons, forums, and other events.

Other Custom Requirements:Including personalized services such as report customization, market research, and recruitment support.

Business Cooperation Discussions,Please contact WeChat: ZuoLeiLeiya

(Note: Name - Institution - Collaboration)

Submission | Become a CreatorPlease contact WeChat: RoseBCI

?Starred & Pinned?

Stay Ahead of Every Breakthrough in Brain-Computer Interfaces

One-click Triple Action: “Share,” “Like,” and “Mark as Read”

Welcome to chat in the comments section