NeuroXess eyes mass BCI production with new unit coming online in H2

Chinese brain-computer interface company NeuroXess is gearing up for what it sees as the next phase of the industry: mass production.

The startup plans to begin operating a 14,300-square-meter manufacturing facility in Jiangxi province in the second half, a move NeuroXess founder and chief scientist Tiger Tao Hu said will help transform BCIs from experimental technologies tested on dozens of patients into products that can be delivered at an industrial scale.

This also reflects the company's dual-core strategy of "research and development in Shanghai, production in Jiangxi", Tao said in an exclusive interview with China Daily.

Such efforts come as BCIs, which create direct communication pathways between the human brain and technology such as a computer or artificial limb, attract growing global attention from investors, governments and technology companies.

While the industry remains dominated by clinical trials and research milestones, companies are increasingly racing to answer a more practical question: Who can manufacture these devices reliably, safely and at scale?

For NeuroXess, the answer lies not only in neuroscience, but also in factories.

The company has completed more than 50 clinical trials and is developing what it calls a fully implanted, fully wireless and fully functional BCI system. The Jiangxi facility is designed to eventually support annual production at the scale of tens of thousands of units.

The company's first patient implanted with the system in late 2025 achieved a brain control speed of 5.2 bits per second after 17 days of training, according to the company, enabling web browsing, gaming and control of external devices.

By early 2026, the participant was independently shopping, writing Spring Festival calligraphy and using digital payment tools through brain control systems.

According to Tao, 2026 has become a pivotal year because three critical pieces are finally coming together.

"The technology, the clinical validation and the manufacturing capabilities have matured at the same time," he said.

As competition intensifies between China and the United States, Tao argues that the discussion often focuses too heavily on individual technologies rather than the broader ecosystem needed to commercialize them.

He rejects the idea that China lags significantly behind the US in BCIs. "The core scientific principles are not fundamentally different," he said.

Tao acknowledged that the US still holds advantages in specialized chips, but argued that China has significant strengths in clinical resources, manufacturing and industrial integration.

BCIs, he said, are not simply chips implanted into the brain, "they are system engineering projects".

The challenge is integrating electrodes, chips, packaging, software, operating systems, wireless communication systems, manufacturing and clinical workflows into a single functioning product.

"A patient doesn't need a world champion in one component," Tao said. "A patient needs a complete system that works."

China's manufacturing ecosystem, he argues, offers a unique advantage. "Technology only matters if patients can actually use it," he said. Tao pointed to China's precision manufacturing supply chains, hospital networks and engineering talent as factors that could help accelerate commercialization.

The Jiangxi factory reflects that strategy. Beyond manufacturing, it will also serve as a rehabilitation and training center. The facility will include simulated streets, subway stations, cafes and convenience stores where patients can practice using BCI systems in realistic environments.

The goal, Tao said, is not simply restoring movement. "The ultimate goal of rehabilitation isn't that a patient can move again," he said. "It's that they have the confidence to go outside again."

In addition, the recent boom in artificial intelligence may prove as important to BCIs as advances in hardware.

Tao describes large AI models not as an add-on to BCIs, but as an "operating system upgrade".

Previously, decoding neural signals often required weeks of calibration for each individual patient. AI models are increasingly making brain signal decoding faster, more accurate and more transferable between patients.

"AI is turning brain computer interfaces from something that must be calibrated for every individual into a technology that can generalize across patients," Tao said.

NeuroXess has already used AI to achieve several notable milestones.

In late 2024, the company conducted what it said was the world's first real-time Chinese language decoding trial. After nine days of training, a participant achieved 71.5 percent accuracy in full-spectrum Chinese syllable decoding without assistance from a language model.

The patient was able to generate spoken Chinese through thought, control a digital avatar and interact directly with AI systems.

In another clinical trial conducted with Shanghai's Huashan Hospital Affiliated to Fudan University in 2025, patients in Shanghai and Fuzhou completed what NeuroXess described as the first long-distance "thought conversation" between two human subjects.

The system translated neural signals into text at speeds exceeding 300 Chinese characters per minute, roughly twice the pace of ordinary speech.

Tao believes such breakthroughs could eventually allow direct interaction between the human brain and AI systems.

"We don't see brain computer interfaces as just medical devices," he said. "We see them as the bridge connecting human intelligence and AI."

chengyu@chinadaily.com.cn