Yunqi Capital and HSG Co-Lead Huachao Shenkong's Pre-A Round, Betting on Non-Invasive AI Brain-Computer Interfaces

As brain-computer interfaces (BCIs) move from concept to commercialization, the real challenge is building stable, deliverable systems that integrate brain reading, brain writing, and AI neural decoding. BCI-Sonics (华超神控) uses low-intensity transcranial focused ultrasound for brain writing, combined with multimodal brain reading via fUS, EEG, fMRI, and PET, plus AI neural decoding, to build a product portfolio for research, clinical, and consumer technology scenarios. The company recently completed its Pre-A funding round, co-led by Yunqi Capital and HSG.

In this edition of Yunqi Partners, we introduce BCI-Sonics — another Yunqi investment in the non-invasive AI BCI space, as we continue tracking the evolution of human-machine endpoints.

---

Yunqi Capital's Investment Thesis:

We are firmly bullish on next-generation convergent solutions built on ultrasound technology and centered on non-invasive closed-loop BCIs. Ultrasound-based BCIs will be the first to break through complex medical scenarios, achieving high-quality deployment in real clinical environments. As non-invasive, high-throughput, high-quality data continues to accumulate, the generalization capabilities of brain-computer models will leap significantly, opening broad commercial prospects and strategic value in consumer applications and next-generation human-computer interaction.

BCI-Sonics possesses rare full-stack capabilities in the industry, spanning foundational R&D in materials, processes, and algorithms through to commercial industrialization. Its core team brings together top talent from leading global medical device, neuroscience, large model, and consumer electronics companies, combining bottom-up innovation with engineering execution. We look forward to the company's products genuinely serving broad populations — enabling safer, more non-invasive BCI technology to move from medicine into daily life, from specialists to universal access.

Recently, non-invasive AI BCI company BCI-Sonics (华超神控) announced the completion of a RMB 200 million Pre-A funding round. The round was co-led by Yunqi Capital and HSG, with follow-on investments from Bilinxing Investment, Oriza Holdings, Xuhui Sci-Tech Investment, and Deshi Investment. Existing shareholders Matrix Partners China and Delian Capital continued to increase their stakes. Xunguang Capital served as the exclusive strategic financial advisor. The proceeds will be used for product R&D, core talent recruitment, clinical research advancement, and AI infrastructure development. Previously, the company completed an angel round led by Matrix Partners China and an angel+ round co-led by Delian Capital and Tao Capital, bringing total cumulative funding to nearly RMB 300 million since founding.

Founded in 2025 and headquartered in Shanghai, BCI-Sonics builds on ultrasound neuromodulation, multimodal brain reading, and AI neural decoding as its core technical pathways. It has structured three product lines around research, clinical, and technology applications, with its main development goal to become China's benchmark enterprise for non-invasive AI BCIs, and on that foundation grow into a Neuro AI company built on brain data.

BCI-Sonics is one of the few innovative non-invasive AI BCI companies in the industry whose milestones lead its valuation. The company's FUS-BCI series of two product generations have completed R&D and obtained corresponding safety reports; its closed-loop ultrasound BCI system is planned to enter clinical trials in Q1 2027; its consumer-facing closed-loop BCI technology product has entered batch validation; and the company is accelerating R&D in multimodal brain reading, Neuro AI platforms, and CNS drug-device combination products, with international major pharmaceutical companies already reached for strategic partnerships in relevant directions.

Non-Invasive AI BCIs: From Technical Pathway to Industry Consensus

Investment logic in the BCI industry is shifting from "evaluating concepts" to "evaluating commercialization." According to IT Juzi data, in the first half of 2026, domestic BCI financing exceeded 60 deals with total value surpassing RMB 7 billion — already exceeding the full-year 2025 figures of 49 deals and RMB 2.723 billion. On the policy front, in June 2025, the Ministry of Industry and Information Technology and six other departments jointly issued a dedicated BCI industry policy; the "15th Five-Year Plan" recommendations listed BCIs as a key future industry to cultivate; by end-June 2026, the National Medical Products Administration released two guidance documents for BCI medical devices; and Hubei, Shanghai, Guangdong, and other regions successively introduced pricing items, insurance coverage, and industrial support policies.

On technical pathways, invasive BCIs have obtained China's first medical device registration certificate, with regulatory paths becoming clearer, but the requirement for craniotomy means they serve only a minority of severe patients. Traditional non-invasive methods such as EEG and transcranial electrical stimulation are limited by spatial resolution and penetration depth, making it difficult to reach deep nuclei. Low-intensity transcranial focused ultrasound (tFUS/LIFU) is currently the only physical pathway that simultaneously offers non-invasiveness, deep reach, and millimeter-scale spatial resolution — making it the route where global capital has concentrated most heavily in the non-invasive direction over the past year: Merge Labs, co-founded by Sam Altman with OpenAI participation, launched with a USD 252 million seed round and in September 2026 signed a multi-year ultrasound chip licensing agreement with Butterfly Network; Nudge, founded by Coinbase co-founder Fred Ehrsam, raised a USD 100 million Series A led by Thrive Capital and Greenoaks, entering via headphone form factor for addiction, chronic pain, and cognitive enhancement.

Since its founding, BCI-Sonics has taken non-invasive AI BCIs as its company主线: using ultrasound to solve the depth and coverage problems of "brain writing," using multimodal neural reading to solve the information dimensionality problem of "brain reading," and using AI neural decoding to close the loop between the two. In June 2026, the company was selected for the "2026 (11th) China's Most Investment-Worthy Companies Top 100" list jointly published by Dongshihui Network and Yabuli Think Tank — one of five BCI companies on the list. This round being co-led by HSG and Yunqi Capital signals that leading domestic institutions have reached consensus on the prospects for this route's commercialization in China.

Brain Writing, Brain Reading, Decoding: The Closed Loop of Ultrasound, Multimodality, and AI

BCI-Sonics' technical system divides into three layers — brain writing, brain reading, and decoding — which form a closed loop among them.

Brain writing relies on ultrasound. The company uses low-intensity transcranial focused ultrasound to penetrate the skull and deliver quantifiable neuromodulation to deep nuclei and superficial cortex, with coverage extending from cortical regions to deep targets unreachable by traditional non-invasive means. The engineering core of transcranial focused ultrasound is phase correction: individual variations in skull thickness, density, and curvature cause acoustic field distortion, energy attenuation, and focal point shift, directly affecting stimulation efficacy rate, dose consistency, and clinical repeatability. BCI-Sonics has built proprietary transcranial phase correction algorithms that can complete individualized calibration in short timeframes; its quasi-commercial device is currently one of the few tFUS/LIFU systems domestically with this capability.

Brain reading relies on multimodality. The company has built a multimodal neural reading system centered on functional ultrasound imaging (fUS) and electroencephalography (EEG), combined with functional magnetic resonance imaging (fMRI) and positron emission tomography (PET), with passive cavitation detection (PCD) for real-time acoustic safety monitoring. Different modalities complement each other in temporal resolution, spatial resolution, and penetration depth, jointly providing high-dimensional brain state inputs for AI decoding.

Decoding relies on AI. Built atop multimodal signals, the company has established AI neural decoding models that identify current brain states, generate individualized stimulation protocols, and continuously optimize parameters based on post-stimulation neural feedback, forming a closed loop of "read — understand — modulate — feedback — re-optimize." Two productized carriers at this layer are: the ultrasound neuromodulation parameter interactive AI platform developed based on closed-loop modulation large models, and the interactive human brain atlas Cortex Atlas released in September 2026 — the former transforming stimulation parameter selection from experience-driven to model-driven, the latter providing structured knowledge of "target — circuit — effect" to support closed-loop stimulation protocol design.

Research, Clinical, Technology: Three Product Lines to Explore, Heal, and Enhance the Brain

BCI-Sonics implements the above technical system through three product lines, serving three categories of users.

The research product line targets neuroscience and brain science research, positioned to explore the brain. In September 2026, at the Chinese Neuroscience Society annual meeting (CNS 2026), the company released its non-invasive closed-loop BCI system for research users, the ultrasound neuromodulation parameter interactive AI platform, and Cortex Atlas — providing a complete toolchain for brain mechanism research from stimulation, through reading, to decoding. In the company's view, mechanism research determines target selection and application depth for clinical and consumer products, making it the upstream of the other two product lines.

The clinical product line targets medical institutions, positioned to heal the brain. Indication coverage spans cortical and deep targets, including pain, Parkinson's disease, Alzheimer's disease, addiction, sleep disorders, and directions such as ultrasound-mediated blood-brain barrier opening (BBB Opening) for drug delivery in brain tumors and neurodegenerative diseases. The company has partnered with multiple top-tier hospitals and clinical PIs to launch multi-center investigator-initiated trials (IITs), while simultaneously advancing medical device registration pathways.

The technology product line targets families and ordinary consumers, positioned to enhance the brain. Product form is wearable BCI terminals, providing non-invasive neural modulation for daily scenarios such as sleep, focus, and mood. The company's goal is for every ordinary person to own their own BCI device.

From Sensors to Neuro AI: Built on Brain Data

In BCI-Sonics' planning, research, clinical, and technology devices share one underlying identity — sensors that collect brain data. Every stimulation and response, every segment of multimodal neural signal, is a sample of how the human brain works. As devices scale across laboratories, hospitals, and homes, the company will accumulate brain data covering different populations, brain states, and intervention conditions, forming a positive flywheel of "data — model — efficacy."

Built on this data, BCI-Sonics' long-term goal is to become a Neuro AI company: first constructing digital twin brains and neural world models capable of predicting individual brain responses to stimulation, then exploring Human AGI with human brain-level consciousness, thinking, and reasoning capabilities. Today's large language models learn from text written by humans; BCI-Sonics hopes future AI will directly learn from human brain activity itself. The company summarizes its current stage as Neuroscience for AI: using neuroscience methods to collect and understand the brain, providing data and foundational principles for next-generation AI. The AI infrastructure portion of this funding round will be directed toward computing power, data platforms, and model R&D.

Team:

Combined Capabilities in Engineering, Clinical, and AI

BCI-Sonics was founded in 2025 by serial entrepreneur Xin Li. Li holds a PhD in biomedical engineering from the joint program of the Chinese Academy of Sciences and Germany's Fraunhofer IGD Institute. He joined GE Healthcare in 2015, rising from life sciences scientist to head of GE Global Research Center China, and previously founded innovative technology and medical companies, with long-standing accumulation in medical imaging, ultrasound, artificial intelligence, and BCI R&D and industrial translation.

The core team comes from Tsinghua University, Imperial College London, Shanghai Jiao Tong University, Zhejiang University, and other domestic and international institutions, covering neuroscience, acoustics, medical imaging, artificial intelligence, materials science, clinical translation, medical device engineering, and commercialization — with complete experience from proof-of-concept prototypes, engineering prototypes, through clinical validation and registration translation. Following this funding round, the company will focus on recruiting core talent in AI neural decoding, clinical registration, and consumer hardware.

Xin Li, founder and CEO of BCI-Sonics, stated: "Over the past year, the BCI industry has completed its switch from 'evaluating concepts' to 'evaluating commercialization.' What investors ask is no longer whether the route is right, but when products enter hospitals and when they enter homes. Our requirements for ourselves are clear: brain writing must reach deep regions and cover the whole brain; brain reading must be multimodal and quantifiable; decoding must rely on AI rather than experience. The three product lines answer three questions: the research line answers how the brain works, the clinical line answers how to treat disease, and the technology line answers how to enter everyone's life. But all devices are fundamentally sensors. When brain data accumulates to sufficient scale and quality, it will become the foundation for next-generation AI. Today's language models learn from text written by humans; we hope future AI can learn from the human brain itself. This is what we mean by Neuroscience for AI. We thank our investors for their trust; this round of funding will be fully invested in building products, talent, clinical capabilities, and AI infrastructure."