Neuralink’s VOICE trial turns intended speech into audible words

The words came through a computer, in a voice that sounded like Terry’s own. Terry, who lives with ALS, was using an experimental brain implant to turn the words he was trying to say into speech that others could hear.

Neuralink shared the demonstration in a September video. As Terry tried to speak, software used in Neuralink’s VOICE trial recognised his brain signals. The software interpreted those signals and turned them into a computer-generated voice that sounded like Terry's own. For people who struggle to speak, this breakthrough offers a deeply personal possibility. It points toward a future where they may regain everyday spoken exchange: speaking to loved ones, answering a question, joining a conversation, or telling someone what they need.

The VOICE trial Terry participated in tested technology designed to help people with severe speech impairment. It is a wireless implant with about 1,024 electrodes across 64 threads. Using these threads, the technology detects brain activity associated with speech production. Then it uses software to interpret those signals as speech sounds and words, before finally playing those words aloud via text-to-speech or a voice clone.

Apart from Terry, Kenneth Shock also participated in a VOICE trial. In March 2026, Neuralink featured him in a demonstration of the technology, turning attempted speech into audible words. He received his implant in January 2026 and is described as the second VOICE participant.

Neuralink’s public materials still frame these VOICE trials as early feasibility work. In late January 2026, the company said it had 21 trial participants across its clinical programmes and that VOICE had recently launched, with an aspirational target of around 140 words per minute for conversational speech. On May 1, 2025, the FDA granted Neuralink a Breakthrough Device Designation for speech technology, a designation intended to expedite development and regulatory review.