Exclusive: A Quantum Computing Unicorn Arrives! Breaking Through the Computational Power Ceiling of Brain-Computer Interfaces

In the rehabilitation training room, a patient equipped with a brain-computer interface (BCI) raised their arm, issuing commands to an intelligent prosthetic limb through thought alone. The command originated from the cerebral cortex, passed through electrodes, was decoded by algorithms, and reached the prosthetic limb within 0.5 seconds. The prosthetic indeed moved, but the brain clearly perceived it as merely a prosthetic device.

A mere 0.5 seconds may seem negligible, yet it is enough for the human body to perceive latency, preventing the emergence of a sense of “embodiment.” If this latency cannot be reduced, prosthetic limbs will remain mere tools, never truly integrating with the body. This scenario epitomizes the lag issues plaguing brain-computer interfaces (BCIs), leaving the industry at an impasse.

A breakthrough in April revealed a potential solution. A paper jointly published by the Shanghai Institute of Microsystem and Information Technology of the Chinese Academy of Sciences and Beijing Bose Quantum Technology Co., Ltd. achieved microsecond-level neural decoding on a real 1,000-qubit specialized quantum computer for the first time. This also laid the groundwork for the current strategic partnership between China’s first brain-computer interface (BCI) special fund—the Tianjin Brain-Computer Interface Heying Haihe Haitai Venture Capital Fund (hereinafter referred to as the “Heying BCI Fund”)—and QBOSON.

Notably, QBOSON recently formally submitted its IPO tutoring registration application to the Beijing Securities Regulatory Bureau, thereby initiating its A-share listing process.

As envisioned by both parties, QBOSON will provide cloud-based dedicated quantum computing power for brain-computer interfaces (BCIs), accelerating the interpretation of deep-seated intentions within the cerebral cortex at specific stages. Within 3–5 years, externally connected dedicated quantum computing hardware is expected to accelerate neural decoding, with potential applications in language and motor decoding scenarios.

So, how will the “water” of quantum computing quench the “thirst” of brain-computer interfaces?

Brain-computer interfaces have long faced the problem of "lag."

The human brain possesses nearly 100 billion neurons, with each instance of cognitive activity corresponding to the synchronous firing of massive neuronal ensembles. Brain-computer interfaces (BCIs) need to acquire these weak and noisy electrical signals and translate them into machine commands via algorithms.

The challenge lies in the fact that these brain signals are not only weak but also contaminated with substantial noise, yet decoding must be performed in real time. During rehabilitation training, even a 0.5-second delay in the smart prosthetic limb’s response to finger-lifting commands can induce a sense of alienation in the brain, making it perpetually difficult for the prosthetic to integrate into the body.

Quantum computing has inherent advantages in solving this problem.

The smallest storage unit of a classical computer is the bit, which can only be 0 or 1. The smallest unit of information in quantum computing is the qubit, which can exist in a superposition of both 0 and 1 states simultaneously before being measured.

Meanwhile, multiple qubits can form strong correlations through quantum entanglement. This "quantum parallelism" feature gives it a potential exponential speedup advantage when dealing with combinatorial explosion problems. For example, 10 qubits can simultaneously encode 2^¹⁰=1024 superimposed states, capable of simultaneously encoding a massive number of computational paths. Although only one result can be obtained through a single measurement, its advantage lies in leveraging the principle of interference to amplify the probability of obtaining the correct solution.

Chen Liang, Vice President of QBOSON, used a more accessible analogy to explain quantum computing: “For instance, in a noisy environment, the human ear automatically locks onto the desired sound, picking it out from the superposition of mixed sound waves. Similarly, the work of quantum computing is to extract the most desired result from the superposition of wave functions.”

He further stated that this characteristic gives quantum computing a natural advantage in processing probabilistic and noisy signals, which aligns precisely with the characteristics of brain signals. However, Chen Liang also emphasized that the role of quantum computing is to serve as a backend computational power engine, rather than merely being a faster conventional computer. It does not directly participate in reading brain signals; instead, it assists in noise reduction, feature extraction, and signal encoding of neural signals.

In Chen Liang’s view, quantum computing power holds promise for addressing several pain points in brain-computer interfaces.

First is the issue of latency. As the number of channels in brain-computer interfaces (BCIs) rapidly increases, real-time decoding latency also rises correspondingly. It is widely acknowledged within the industry that closed-loop BCI latency typically needs to be reduced to below 50 milliseconds to prevent the natural body from experiencing a significant sense of rejection toward external devices such as smart prosthetic hands. Quantum computing can accelerate neural signal processing and decoding in BCIs, thereby alleviating the computational burden and response latency caused by the increase in channel count.

Secondly, there is the issue of signal interpretation. As previously mentioned, brain signals are weak and fraught with noise, whereas quantum computing excels at processing probabilistic and ambiguous information, enabling the precise capture of faint signals from background noise. This will provide significant benefits for the treatment of brain disorders, such as allowing implantable deep brain stimulators to target only those brain regions exhibiting abnormal electrical discharges, thereby sparing healthy tissue.

Finally, predicting disease onset. Taking epilepsy as an example, subtle abnormal fluctuations occur in the brain prior to a seizure. Quantum computers, which excel at combinatorial optimization, can assist in identifying these relevant changes and forecasting seizure trends, thereby securing a valuable time window for intervention.

How do brain-computer interface (BCI) companies view these advantages of quantum computing? VCBeat also engaged in discussions with industry insiders.

Gnosis Healthineer primarily focuses on big data and foundational model platforms for brain-computer interfaces (BCIs). According to its leadership, quantum computing can help this system overcome the computational bottlenecks encountered when advancing toward high-density, real-time closed-loop BCIs. Particularly as the dimensionality of neural signals continues to increase, the potential of quantum computing in combinatorial optimization and high-dimensional feature representation offers new computational pathways for real-time neural decoding, online recalibration, and the discovery of latent patterns in complex brain disorders. This advancement will propel BCIs from merely “reading” the brain to a combined paradigm of “reading and intervening.”

WABO Technology has been deeply engaged in the field of neural intent model construction. Its leadership believes that for brain-computer interface (BCI) “model companies” like WABO, quantum computing holds the promise of delivering acceleration unattainable by classical computing power in addressing challenges such as low latency, cross-subject generalization, reduced calibration needs, and large-scale neuronal simulation—all without altering the core logical framework of “neural data input → intent modeling → physical control → data feedback.” This advancement not only raises the performance ceiling for deploying intent models in the physical world but also aids in uncovering the underlying mechanisms of the brain’s intent layer and even the origins of consciousness.

Beyond its application value, capital is also paying attention to this sector. Zhang Yutao, a partner at the Heying Brain-Computer Interface Fund, stated that Heying’s foray into quantum computing was not a spur-of-the-moment decision: “The Heying Brain-Computer Interface Fund has always regarded brain cognition and brain decoding as key investment areas. Multimodal and large-model brain-computer interface applications heavily rely on computational power for decoding, real-time optimization, and small-sample classification. This bottleneck in computational capacity is precisely where quantum computing can offer improvements.”

Is the Logic of Using Quantum Computing Power to Solve Brain-Computer Interface Problems Valid? In Fact, Prior to This Collaboration, Preliminary Answers Had Already Been Provided by Real-Machine Verification.

In May this year, QBOSON released its new generation of specialized quantum computer, the “Yuliang Shanhai 1000.” Almost simultaneously, the Shanghai Institute of Microsystem and Information Technology of the Chinese Academy of Sciences jointly published a paper with QBOSON, marking the first successful validation of neural signal decoding on a real-world 1,000-qubit specialized quantum computer. This remains the only such validation of neural signal encoding and decoding performed on actual hardware in China to date.

This successful validation has also served as a major boost of confidence for the entire industry. Chen Liang explained to VCBeat that quantum computing is divided into general-purpose and specialized quantum computing. The industry as a whole remains in the Noisy Intermediate-Scale Quantum (NISQ) era, with general-purpose fault-tolerant quantum computing likely still 5–10 years away from mature applications. However, specialized quantum computing has already entered the practical validation stage in certain scenarios, indicating that the entire industry is gradually transitioning from laboratory research to industrial validation.

Quantum computing encompasses various technological approaches, including superconducting circuits, ion traps, and photonic quantum systems, each with its own advantages and disadvantages. Why is QBOSON’s approach currently considered the most suitable? VCBeat provides an explanation.

First is the technical roadmap. QBOSON is one of the few dedicated quantum computing companies in China, having built its product line around coherent optical quantum computers.

Chen Liang explained to VCBeat that photons inherently possess quantum properties and, unlike superconducting or ion trap systems, do not impose stringent requirements on supporting infrastructure for preparation. “For instance, the superconducting approach requires massive ultra-low-temperature dilution refrigerators. In contrast, certain core components of QBOSON’s 1,000-qubit specialized quantum computer can operate at room temperature. The system is available for use no less than 12 hours per day, with a mean time between failures (MTBF) of no less than 1,000 hours.”

Next is the validation of integrated hardware and software. Zhang Yutao introduced that the current industry consensus is that a hybrid architecture comprising “classical computing power for front-end preprocessing (such as GPUs and ASICs) plus an externally connected dedicated quantum accelerator” is the most feasible approach. QBOSON’s dedicated quantum computer offers advantages including stable operation at room temperature and atmospheric pressure, low inference latency, and compatibility with sparse spike signals, making it well-suited for online real-time inference.

On the software front, Chen Liang stated that QBOSON already offers toolchains and SDKs compatible with Python, significantly lowering the barrier to entry: “There is no programming burden; even an undergraduate student can get started. The real challenge lies in formulating problems into a format solvable by specialized quantum computers.”

More critically, QBOSON delivers an integrated hardware-software solution rather than relying solely on digital simulation. Zhang Yutao believes this is crucial in the medical field: "Many brain-computer interface (BCI) innovations remain at the stage of digital simulation, but regulators have low tolerance for purely simulated models and high concern regarding algorithmic black boxes. Purely digital simulations face significant hurdles in pursuing medical regulatory pathways. QBOSON has completed closed-loop validation of neural decoding and inference using real, specialized quantum computing hardware. A closed loop has been established among its hardware, algorithms, and datasets, and joint validation with in vivo datasets has already been conducted with several BCI hardware teams. Furthermore, the company independently develops the full technology stack—including photonic quantum chips, optical circuit systems, complete machines, and software—ensuring there are no chokepoints or dependencies on external suppliers."

Third, strategic ecosystem positioning. The Heying Brain-Computer Interface (BCI) Fund’s ecosystem resources already cover more than 360 BCI companies, facilitating the promotion and diffusion of quantum computing power within the BCI industry. QBOSON provides dedicated quantum computing power, distributed cloud services, and an AI toolchain. Both parties ultimately aim to pioneer the first hybrid quantum-classical computing solution for brain-computer signal processing in China.

The technical solution is led by both parties, who have joined forces with brain-computer interface (BCI) companies to conduct technical validation. Meanwhile, commercial validation has been incorporated into the design from the outset. Zhang Yutao stated that multiple BCI enterprises, including Gnosis Healthineer, which specializes in big data and foundational model platforms for brain-computer interfaces, and WABO Technology, which focuses on building neural intent models, have already participated in the validation process. All parties are working to build an ecosystem that integrates quantum computing and BCI technologies for commercial implementation, and are actively establishing substantial infrastructure for real-world deployment.

As two emerging industries under the 15th Five-Year Plan, brain-computer interface (BCI) technology and quantum technology are resonating with and reinforcing each other on the path of human civilization’s evolution. The intricate entanglements and vast mysteries surrounding the 86 billion neurons, the brain’s intention layer, the origins of consciousness, and their relationship with human civilization may require more powerful computational capabilities to unravel.

Regarding the future of combining quantum computing with brain-computer interfaces, Chen Liang and Zhang Yutao provided complementary timelines.

In the short term, quantum computing power will primarily serve fundamental research in brain science. Chen Liang believes that the entry point should focus on research institutions and clinical studies, such as neuronal network modeling and simulation, and research into the mechanisms of diseases like epilepsy. He admits that current devices still provide services via the cloud, but since cloud-based solutions still suffer from latency, local deployment will be the ultimate goal in the future.

Zhang Yutao offered a more specific assessment: “In the mid-term horizon of three to five years, dedicated quantum computing hardware for external neural decoding acceleration can be practically implemented, particularly in the domains of language and motor decoding. Examples include real-time control of high-degree-of-freedom prosthetic limbs, EEG-to-speech conversion for aphasic patients, and rapid calibration with closed-loop modulation following stroke or spinal cord injury. Subsequently, quantum machine learning and classical deep learning may converge to address critical challenges such as cross-subject generalization and few-shot calibration. In the more distant future, the field will point toward large-scale neuronal cluster dynamics simulation, multi-brain-region modeling, and even edge deployment—essentially transforming quantum computing power into portable, wearable devices.”

As for scenarios circulating online, such as intracranial implantation of quantum chips and quantum brain-computer interfaces capable of reading consciousness, Zhang Yutao candidly stated that “such possibilities are not currently foreseeable.”

Regarding application focus, Zhang Yutao believes that quantum computing holds greater value for non-invasive brain-computer interfaces (BCIs) than for invasive ones: “Quantum computing addresses backend algorithms and computational acceleration, but it cannot improve electrodes, implantation procedures, or biocompatibility. Currently, non-invasive BCIs face significant signal-to-noise ratio challenges, and with the more rapid development of multimodal big data, the demand for quantum computing will be even greater.”

However, both experts also acknowledged that both brain-computer interfaces and quantum computing are currently in their early stages and still face numerous challenges.

Cross-Domain Interfaces and Data Flow Present the First Challenge. The latency of closed-loop brain-computer interfaces must be reduced to below 50 ms. However, both neural signals and quantum hardware inherently exhibit signal noise. Although quantum hardware itself can achieve microsecond-level performance, the entire process—encompassing acquisition, preprocessing, encoding, transmission, and feedback—accumulates significant latency.

On the hardware front, quantum computing remains immature, presenting significant room for development. On the software side, mapping neural tasks to quantum models remains a challenge; there is a lack of native quantum algorithm libraries, and quantum acceleration benefits only specific optimization tasks.

Brain-computer interface terminals also face their own challenges, including issues with electrodes, biocompatibility, and implant stability.

Meanwhile, challenges remain in engineering and clinical translation. Currently, talent spanning both quantum computing and brain-computer interfaces (BCIs) is extremely scarce. Therefore, enabling BCI developers to easily access quantum computing power through software development kits (SDKs) and toolchains represents a more practical approach at present.

Chen Liang outlined several pathways to address these challenges: “We must adhere to a hybrid approach combining quantum and classical computing, leveraging quantum computing for localized enhancement and acceleration through iterative, stepwise progress. Simultaneously, we should establish collaborative teams that bring together quantum computing, brain-computer interface engineering, and clinical institutions, starting with small-scale prototype validation. Furthermore, we need to advance the engineering of hardware to improve stability and control costs.”

In 2026, brain-computer interfaces (BCIs) are widely regarded as entering their inaugural year of commercialization. According to McKinsey’s estimates, the potential market size for serious medical applications of BCIs could reach $15–85 billion. However, despite being on the verge of explosive growth, BCI technology still faces shortcomings in underlying computational power. Meanwhile, the latency issue confronting BCIs has become even more urgent.

The entry of quantum computing, another frontier technology, may hold the promise of resolving the dilemmas facing brain-computer interfaces (BCIs). Only when decoding latency is compressed to below the hundred-millisecond level can exoskeletons and prosthetic limbs truly become integrated extensions of the body; only when computational power enables real-time analysis of massive-scale joint neuronal firing can BCIs truly comprehend the brain’s intentions.