Sabi raises $50 million seed, led by Khosla, for brain-to-AI interface in a cap

Sabi, a start-up based in Palo Alto, California, said on Friday, October 9, 2026, that it has raised a $50 million seed round led by Khosla Ventures to build a brain-to-AI interface that people can wear every day. Its first product is a cap with sensors hidden inside that read the brain's electrical signals through the hair, without gel, surgery or direct contact with the scalp, and turn imagined speech into commands for AI software. Accel, Initialized Capital, DST Global and former OpenAI product chief Kevin Weil also invested. The company says it will show an early prototype at CES 2027.

Brain-computer interfaces have mostly meant one of two things: implants placed in the brain by surgery, which can be precise but are risky and meant for patients, or EEG headsets with wet electrodes, which are safe but awkward and noisy. Sabi is betting that better sensors, a low-power chip and modern AI can make a non-invasive device good enough for everyday use, hidden in something as ordinary as a baseball cap.

The pitch ties neatly to the AI boom. Today people talk to AI tools by typing or speaking. A device that could turn the words you imagine saying into prompts would add a silent, hands-free way to use AI. The company says its model can read imagined speech and even anticipate a few keystrokes before a user types them. If that works reliably, it could help people with disabilities that limit speech or movement as well as ordinary users.

The caution is significant. Reading clear signals through hair and skull is hard, and results can drift from one session to the next. The company's speed and accuracy claims come from its own tests, and reports differ on basic specifications such as sensor size and count. A $50 million seed round, very large for a first round, shows investor appetite, but the device has yet to be shown publicly.

Privacy is the other big question. A device that reads brain signals collects some of the most sensitive data there is. How that data is stored, used and protected will matter as much as how well the cap works, for regulators in the US and in markets such as India.

For Indian engineers and researchers in neurotechnology and AI hardware, the round signals that wearable brain interfaces are moving from labs to venture-backed products.

Short-term: Sabi will spend on hardware and on training its brain-signal AI model, aiming to demonstrate a working prototype at CES 2027.

Long-term: If non-invasive brain interfaces become accurate and comfortable, they could become a new way to control computers and AI, raising new questions about neural data privacy.

Who is affected: AI and wearable-device companies, people with speech or mobility impairments, privacy regulators, and investors in neurotechnology.

Questions and Answers