(Yicai) June 2 -- Brain-computer interface will become the translation layer between machine intelligence and the neural language of the human brain, enabling direct mind-to-machine communication without the need for spoken language, according to the founder of leading Chinese brain-computer interface firm Gestala.

AI is central to overcoming the limitations of ultrasound-derived neural data, Phoenix Peng, who is also Gestala's chief executive officer, said at the Beyond Expo 2026 in Macao on May 30.

Ultrasound produces large-scale but noisy hemodynamic signals across broad brain regions, the kind of high-dimensional data set where machine learning has repeatedly outperformed conventional signal processing, he explained. “Without AI, scalable whole-brain decoding would be extremely difficult.”

Gestala, very similarly to OpenAI-backed Merge Labs, is working toward neural-decoding systems applicable across multiple users rather than tailored to individual patients, Peng noted. Gestala's product, however, is considerably more grounded.

Clinical data show that Gestala's magnetic resonance imaging-guided ultrasound device can reduce chronic pain by 50 percent on average after a single session, with effects lasting one to two weeks.

About 20 percent of the global adult population suffers from chronic pain, and current pharmacological options carry well-documented dependency risks. Gestala has not publicly disclosed treatment pricing.

The company plans to file for the National Medical Products Administration's Class III registration by the end of the year, with production of a first-generation device slated for the third quarter, Peng revealed.

NMPA Class III reviews for novel medical devices can take several years, depending on clinical data requirements. However, Peng said that the company hopes to hit the market within one to two years of filing.

The scientific premise behind this product is functional ultrasound imaging, a technique that detects blood flow changes in the brain as a proxy for neural activity.

This approach avoids surgical implantation but generally provides lower spatial resolution than electrode-based systems. Blood flow dynamics lag behind electrical neural signals by seconds, and the skull scatters sound waves before they reach neural tissue. Nevertheless, ultrasonic BCI supports whole-brain signal recording and can effectively access deep brain regions beyond the reach of non-invasive electrical BCI. In comparison, electrode-based systems only capture limited, localized superficial neural signals.

"The two biggest challenges are skull-induced attenuation and distortion, as well as accurately decoding neural activity from complex hemodynamic signals," Peng said.

Independent researchers are cautious about the pace of progress. When Merge Labs launched its ultrasound BCI program earlier this year, Nature reported that researchers said the technology is still at an early stage.

The only published demonstration of functional ultrasound imaging in an awake human required a surgically implanted acoustically transparent cranial window, a significant asterisk on the "non-invasive" claim. Peng acknowledged that Gestala's pathway may also require semi-invasive configurations in certain cases.

The Business Case

There is a huge difference in scale between Gestala and Merge Labs. The former raised USD21.6 million at a valuation of USD100 million to USD200 million, with 25 employees and one production facility in Chengdu, which was just completed in May. The latter has already raised USD252 million.

Peng believes the gap is withstandable, given China's engineering execution speed, hospital network density for clinical recruitment, and the easier regulatory pathway for non-invasive devices, which requires less safety evidence. "Non-invasive ultrasound offers advantages in safety, patient acceptance, and regulatory pathways," he noted.

Gestala's angel round was oversubscribed, with investors including Guosheng Capital, Dalton Venture, Gobi Partners, and Fourier Intelligence.

The technological generation gap between China and the US in electrical BCIs has narrowed from eight years to four years, Peng said, adding that in ultrasound, where clinical and engineering studies only accelerated from 2023, China is only about 18 months behind. BCI appeared in China's national government work report for the first time this year.

"The biggest risk is underestimating both the technical complexity and the time required for rigorous clinical validation," Peng noted.

Editor: Futura Costaglione