Nirva raises $33M total for always-on pendant that listens only to its wearer
Nirva, a startup founded in 2025 by ex-Meta Reality Labs employees Wei Lyu and Jason Chen with Hanying Hu, announced a $25 million pre-Series A round, bringing total funding to $33 million (earlier backers include Lightspeed and South Park Commons). Its first product is a roughly 8 g pendant, also wearable as a bracelet, priced at $249, with up to three days of battery life. Orders open in November, with shipping to early users by the end of 2026. This is business and product news with no direct threat to defenders, though it touches on privacy design in ambient AI wearables.
Detalhes técnicos
- Users record voice snippets during onboarding; the device builds a voice model and filters out other people's speech, so it records only the wearer.
- By default, audio and transcripts are not retained after processing.
- Before discarding audio, the device tracks tone, pitch, volume and speed to infer emotion-related insights, which are surfaced in the companion app.
- The app infers personality patterns, people around the user and places visited.
- Competitors named include Taya, Vocci, Sandbar Ring, Friend, Bee (Amazon) and Ambi.
Linha do tempo
-
Nirva raises $33M total for always-on pendant that listens only to its wearer
ler a análise desta fonte ocultar
Nirva, a startup founded in 2025 by ex-Meta Reality Labs employees Wei Lyu and Jason Chen with Hanying Hu, announced a $25 million pre-Series A round, bringing total funding to $33 million (earlier backers include Lightspeed and South Park Commons). Its first product is a roughly 8 g pendant, also wearable as a bracelet, priced at $249, with up to three days of battery life. Orders open in November, with shipping to early users by the end of 2026. This is business and product news with no direct threat to defenders, though it touches on privacy design in ambient AI wearables.
Technical Details
- Users record voice snippets during onboarding; the device builds a voice model and filters out other people's speech, so it records only the wearer.
- By default, audio and transcripts are not retained after processing.
- Before discarding audio, the device tracks tone, pitch, volume and speed to infer emotion-related insights, which are surfaced in the companion app.
- The app infers personality patterns, people around the user and places visited.
- Competitors named include Taya, Vocci, Sandbar Ring, Friend, Bee (Amazon) and Ambi.