A drone controlled purely by thought, without a single manual command or voice prompt, once seemed like a scene reserved for sci-fi blockbusters like Avatar. Yet, this is now a reality being demonstrated by a Chengdu-based company that is barely a year old.
On August 21st, Zhang Xingzhi, founder of brain-computer interface (BCI) company Wabo Tech, appeared on XWBank's "AI Constructs" live stream, engaging in a deep dialogue with Chen Ran, Deputy Editor-in-Chief of Xinhua Finance, on the technical routes, commercial viability, and industry ethics of the BCI sector. Years after graduating from Tsinghua University, this post-90s founder has charted a deliberately unconventional path, training an "intent large model" on neural data from millions of ordinary people to enable machines to truly "understand" human intent.
From Sci-Fi Seed to Chengdu Innovation
In the film Avatar, a paraplegic protagonist remotely controls his avatar on Pandora through sheer willpower—a scene that planted the seeds of inspiration in a young Zhang Xingzhi and forged his lifelong connection to brain-computer interfaces. Starting in his undergraduate years at Tianjin University, Zhang explored "human-computer interaction," joining a lab in his sophomore year to work on national projects. By his senior year, his thesis involved operating a computer via somatosensory devices, sparking his vision of controlling the digital world with the mind. He subsequently pursued a master's in Engineering Management and a doctorate in Innovation Leadership Engineering at Tsinghua University, focusing consistently on human-computer interaction, BCI, and AI.
Before venturing into entrepreneurship, Zhang spent years embedded in the tech industry, holding roles at major corporations like SAP, Lenovo, and Tencent. He also served as a member of the Ministry of Industry and Information Technology's Metaverse Standards Committee and as Vice President of an A-share listed company, heading its research institute. He authored Industrial Metaverse 3.0 and led the drafting of national standards, deeply involved in industrial and academic collaboration. These experiences drove his focus on a critical question: how can cutting-edge technology evolve from a demonstration into a deliverable, sustainable, and iterative system?
In 2025, BCI was officially incorporated into China's "15th Five-Year Plan" as a key future industry. By March of the same year, Wabo Tech was registered in Chengdu, with operations officially commencing in November. The choice of location was strategic. Zhang noted that the high-tech zone's comprehensive innovation support policies, combined with the agglomeration of chip, AI, and biomedical industries, perfectly suit BCI's interdisciplinary nature. Initially, the company adopted a lean operational strategy, purchasing second-hand hardware, leveraging AI tools to optimize processes, and keeping the team small to maximize efficiency and minimize costs.
To the team's surprise, capital market feedback was far more enthusiastic than expected. Between March and June 2026, within a span of just three months, Wabo Tech secured seed, seed+, and angel funding rounds. The seed round was exclusively invested by Chengdu High-tech Zone's Ceyuan Capital. The angel round, disclosed on July 29th, attracted prominent investors including Vertex Ventures (under Singapore's Temasek and an early backer of Unitree Robotics), Hongtai Fund, Chengdu Future Industry Fund under Chengdu Science and Technology Innovation Investment Group, and Huifengda Capital.
Charting a Non-Invasive Path, Diverging from Mainstream Competition
Elon Musk's Neuralink is an unavoidable benchmark in the BCI field, yet Wabo Tech's approach is fundamentally different. Zhang believes that compared to invasive methods, non-invasive solutions offer a significant advantage due to their lower barriers to entry in everyday consumer applications. Wabo Tech's technology is entirely non-invasive, requiring no surgery. Their devices simply attach to the body's surface to collect EEG signals, making them suitable for daily life.
"Many people mistakenly believe brain-computer interfaces are 'mind-reading,' which is the industry's biggest misconception," Zhang stated during the live stream. He clarified that current technology is far from being able to read a person's private thoughts. Wabo Tech's core technology decodes only specific, controllable intentions—such as choices, confirmations, corrections, and stops—within particular scenarios. "Rather than spying on the mind, we are creating a more natural, seamless, and convenient new entry point for human-computer interaction," he added.
To address the common weaknesses of non-invasive BCI—weak signals and susceptibility to environmental noise—Wabo Tech has adopted a "hardware as entry, model as core" strategy. They have implemented a single-channel, conductive-gel-free EEG recognition solution, enabling devices to be worn and used instantly with no perceptible sensation. Simultaneously, they leverage vast amounts of real-world civilian data to train their AI algorithms, using algorithmic strength to compensate for hardware signal-to-noise limitations and ensure robustness in complex daily environments.
Currently, Wabo Tech's proprietary HAOD R&D and operations system covers the entire process, from self-developed hardware and algorithm noise reduction to EEG command adaptation and data iteration. On the hardware front, they've optimized electrode materials and successfully integrated a domestic AFE analog front-end chip. Their algorithms include a mature multimodal EEG and EMG noise reduction and cleaning model. At the system level, they've achieved intent-based control for a variety of devices, including drones, robotic dogs, smart home systems, and mech suits.
The true barrier in BCI technology lies in data. Unlike the clean, controlled data from small patient cohorts in lab settings, Wabo Tech—while strictly adhering to user consent protocols and industry ethics—collects EEG data from the general population through consumer devices like headphones, lightweight headbands, and skin patches. This happens in real-world scenarios such as brainwave painting, mind-controlled devices, and emotional interaction. This data, inherently containing environmental noise, trains models with greater robustness, making them directly applicable to everyday consumer situations. The company has already launched its proprietary Neural Intent Data Platform (NIDP), which has completed compatibility testing with Ascend AI processors and received Ascend technical certification, thereby establishing a full industry chain from neural signal acquisition, data processing, and model training to edge inference and physical device control.
Products Beyond the Endpoint: How Real-World Data Fuels the Intent Model
Wabo is focusing on two lightweight terminals for its productization efforts. "Lingxi" (WABO Link) is designed for spatial computing and embodied device interaction, serving as the neural intent entry point and device interface in pre-defined tasks. "Dream Weaver" (WABO Dreamer) targets sleep observation, meditation training, and mental wellness management. The long-term value of these terminals lies not just in performing a single function but in establishing user relationships, gathering scenario feedback, and, with clear authorization, accumulating trainable data for the models.
Based on these two core products, Wabo Tech has entered three major commercial scenarios. The first is brainwave painting. Their devices capture and quantify multi-dimensional brain states like focus, emotional fluctuation, excitement, and low mood in real-time. In deep collaboration with research institutions affiliated with West China Hospital, the system generates a unique emotional painting based on a user's 3-5 minutes of live EEG data, translating abstract mental states into tangible visual art.
The second scenario involves mental health and cognitive assistance screening. Leveraging objective and precise EEG data, the company conducts specialized auxiliary screening research for common conditions such as depression, anxiety, early cognitive decline in Alzheimer's disease, and ADHD-related attention deficits. Compared to traditional subjective self-assessment scales, EEG data provides an objective physiological reference, offering a valuable complement to standard questionnaires.
The third scenario is scientific talent assessment. By using quantifiable EEG-based emotional and cognitive evaluation models to replace traditional subjective self-assessment tools like MBTI, Wabo aims to achieve more scientific and precise person-job matching. These solutions are already in pilot use at several central state-owned enterprises. For example, the data can accurately identify traits: individuals with high emotional volatility and flexible thinking are better suited for marketing and creative roles, while those with high empathy and emotional perception are more appropriate for human resources and service-oriented positions.
Upholding Technical Ethics: Human Intent as the Ultimate Control
Each step forward in brain-computer interaction brings sharper scrutiny of ethical boundaries. Wabo Tech, however, adheres firmly to its ethical principles. The company has built an "ethical gate" calibration at the signal input stage, clearly defines capability boundaries on its devices (excluding support for any lethal or harmful applications), mandates informed consent agreements before use, and implements real-time filtering of risky commands via on-device inference.
For Zhang, certain technical bottom lines are non-negotiable: "The physical operation authority of embodied intelligent devices must always remain under human intent control. The core of brain-computer technology is to strengthen human control over intelligent devices, not to cede decision-making power to machines."
On the team and ecosystem front, Wabo brings together researchers from Tsinghua, Tianjin University, and other top institutions with hands-on engineering talent. They have brought on a Chief Scientist from a National 973 Project and a professor from the Shenzhen Institute of Advanced Technology, CAS, as core scientific support. They have also established strategic partnerships with Tianjin University and Fudan University. The team's average age is around 28-29.
Despite its youth, the company is already making its mark on the national stage. On July 31st, its project, "New Generation Brain-Computer Interface Terminal + Brain-Controlled Everything' Based on Ascend Computing Power," won the second prize in the technology category at the "Create the Future" National Entrepreneurship Competition.
This dialogue was part of XWBank's "AI Constructs" series, which focuses on the industrial application of artificial intelligence and frontier technologies. Through company visits and expert discussions, the program showcases the journey of local tech enterprises from R&D to industry implementation. XWBank is also continuously exploring financial services and industry connections to support technology-driven SMEs.
The growth of hard-tech companies from lab to industry is inseparable from financial empowerment and industrial companionship. As a new-generation digital-native bank, XWBank, through its "AI Constructs" program, brings Wabo Tech's achievements and value to the public. This serves both as a window into cutting-edge BCI technology and as a vivid example of how fintech can accompany the growth of local tech innovators. Leveraging its advantages in flexibility, efficiency, a diverse ecosystem, and technological synergy, XWBank provides foundational financial services while also delivering value-added support such as brand IP empowerment and help in amplifying the voice of tech innovation to support the growth of local hard-tech enterprises.