Gu Jie, founder and CEO of Chinese robotics firm Fourier, is pitching what he calls a near‑term leap in rehabilitative care: robots that do not merely follow instructions but read a patient’s intention and assist movement in the precise time window when neural intent is strongest. Far from a distant science‑fiction scenario, Fourier says the key ingredient—brain‑computer interface (BCI) hardware that is portable and mass‑producible—has matured enough for clinical deployment within one to two years.
The company is combining its embodied‑intelligence robots and force‑feedback exoskeletons with BCI sensors to create a closed loop that begins with brain activity. The system, Fourier argues, will detect whether a patient is actively attempting a movement, measure the timing and strength of that intent, and then supply targeted physical assistance so that therapy trains the pathway from central command to peripheral action. Fourier unveiled an upgrade to its “embodied intelligence rehabilitation hub” at a recent conference and outlined a 1+3+X ecosystem strategy—one core self‑operated care scenario, three partner scenes (interactive guidance, industrial enablement, research), and a variable set of ancillary applications.
The technical case rests on two trends. First, decades of BCI research have been turbocharged by recent AI advances, yielding better signal processing, interpretation of weak neural signals, and swift translation into motor commands. Second, China’s device manufacturing ecosystem has pushed BCI hardware toward smaller, cheaper and more portable form factors. Fourier cites progression from conventional EEG to near‑infrared spectroscopy (NIRS) and anticipates ultrasound approaches as additional options.
Fourier’s confidence is bolstered by its internal track record: the company began researching BCI‑driven exoskeletons in 2017 and has demonstrated an EEG‑controlled walking exoskeleton demo. It already supplies robots to more than 2,000 hospitals and institutions across 40 countries, and sees a phased market evolution—from hospital rehabilitation, to care homes, and ultimately into households. If the final domestic phase is realised in the coming five to ten years, Fourier estimates a market of trillion‑yuan scale.
The firm’s roadmap is pragmatic: focus on maturing a reliable hardware platform, validate it in target scenarios, then scale by sharing modules and standards with partners. Fourier describes the next two years as a “1→10→100” process—moving from single validated products to broad platform adoption—and says it will continue heavy R&D investment while deepening collaborations to turn technical advances into measurable commercial value.
Beyond hardware, Fourier is candid about data. Building robots that generalise requires high‑quality datasets that mix publicly available multimodal sources, first‑person human interaction recordings, and smaller but high‑value robot‑collected samples. The company argues that quantity alone is insufficient; diversity of tasks and inclusion of both successes and failures are essential for models that can reason and act in real settings.
The claim that clinically usable BCI‑robot systems are imminent raises several practical questions. Clinical validation, regulatory clearance, long‑term safety and robustness to false‑positive intention detections remain non‑trivial hurdles. Cost, reimbursement, integration with existing rehabilitation workflows, and patient privacy protections will determine whether pilot demos scale into routine therapy. Moreover, ethical safeguards around neurodata and informed consent must keep pace with technical deployment.
Nonetheless, Fourier’s announcement reflects a broader inflection in robotics and healthcare. Governments and industry in China are aligning around assistive robotics and neurotechnology for aging populations; portable BCIs and embodied AI are converging into applications that were once speculative. If Fourier and its peers can clear the clinical and regulatory bar, the result could be a meaningful expansion of rehabilitative capacity and new commercial opportunities for integrated hardware‑software providers.