The Subvocal Company, a San Francisco-based wearables startup co-founded by Aarush Yadav and Arnav Kumar, has raised $1.3 million (about ₹12 crore) in pre-seed funding from Y Combinator and Afore Capital to build a pair of glasses that let people “talk” to their computers without making a sound. Yadav announced the round on X on 30 October 2026, and the company has opened reservations for its first units.

The pitch is simple. Typing is slow, and speaking to a computer in an open office, a café or on a train is awkward for the user and distracting for everyone nearby. Subvocal wants to give users the speed of voice input without the voice.

How the glasses read silent speech

Instead of a microphone or a camera, the glasses send low-power radio waves through the wearer’s face and measure how those signals change as the facial and speech muscles move while the user silently mouths words. The company’s AI models then decode those signal patterns directly into text in real time.

According to the founders, their prototype can already decode open-ended speech at more than 200 words per minute with a reasonably low error rate, even with a small training dataset. The articulation can be clear or subtle enough that someone sitting across the table barely notices. Subvocal says the device does not use cameras or lip-reading, does not listen to the user’s voice and is therefore not affected by background noise. The company has not disclosed its full sensing architecture, citing competitive reasons.

Importantly, the glasses do not read thoughts. Users still have to deliberately form the words with their mouth, just without producing sound.

From hackathon to Y Combinator

The founders began work on the idea roughly four months ago after building a first version at a hackathon. An early, cruder build won Y Combinator’s G1 product showcase. Since then, the team has moved from a table of electronics with antennas taped to the face to a semi-portable board paired with glasses, and says it is close to shrinking the hardware enough to fit entirely inside the frame.

The next step is data. Over the coming months, Subvocal plans to collect thousands of hours of radio-frequency data across a broad population to pretrain a model that works for almost anyone after a short calibration.

Pricing and availability

Reservations are open in the US. The glasses carry a listed price of $699, with early buyers able to reserve a unit for $499 through a refundable $199 deposit. Prescription lenses cost extra. The company currently expects to begin shipping in the first quarter of 2027, subject to regulatory approval, and has cautioned that final specifications may change.

The 200-plus words-per-minute figure comes from the company’s own prototypes and has not yet been independently verified. How well the system copes with different accents, unfamiliar vocabulary and long dictation sessions will be the real test once units reach users.

Why it matters

As AI agents take on more of the actual work, the human job is shifting towards giving detailed instructions, and voice has become one of the fastest ways to do that, as the popularity of AI dictation tools shows. Subvocal is betting that silent speech can become the next input layer for AI, letting people write emails, prompts, notes, messages and even code at the speed of speech without disturbing anyone around them.

For India’s growing community of AI builders and founders, the story is also a reminder that the next frontier in AI is not only bigger models, but new hardware interfaces that change how people interact with them.

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter.

"Exciting news! Elets technomedia is now on WhatsApp Channels Subscribe today by clicking the link and stay updated with the latest insights!" Click here!