All stories
Ambi ambient AI Launches Developing

Ambi Discloses $26M for an AI That Can Listen Through the Day

The startup is opening early access to apps that can gather context across Apple devices, while its optional ambient capture raises a bystander-disclosure question.

Sources (2)
Why now TechCrunch reported Ambi's $26 million raised across multiple rounds on October 12, when the company also described early access to its iPhone, Mac and Apple Watch apps. This is not a newly completed single funding round or a general-availability launch.

An assistant that can work while a conversation continues

Ambi has disclosed $26 million raised across multiple funding rounds, TechCrunch reported on October 12. The startup, led by former Plaud executive Peter Mo, is opening early access to apps for iPhone, Mac and Apple Watch. Users can choose to let them gather context through the day rather than starting each task with a fresh prompt. Ambi says its agent can surface information and begin follow-up work while a meeting is still underway, then carry context between devices. That is the company’s product pitch, not an independently measured productivity result.

TechCrunch says Luminous Ventures and device manufacturer Huaqin led the funding rounds. The $26 million is a total across rounds, not the size of one newly closed round. Ambi’s English site currently invites people to a waitlist and says its hardware will be available early next year. A planned device with a camera, microphones, speaker and screen remains a future product, according to the report.

Listening has a second audience

Ambi told TechCrunch ambient capture is off by default and can be disabled at any time. The company says it deletes audio after transcription and encrypts what remains; its site describes those practices but does not provide an independent security audit. TechCrunch points to a separate disclosure problem: other people in a conversation may not know an app is capturing them. Mo acknowledged that users must tell others when they enable the feature. The launch therefore tests both demand for continuous context and whether users can make that capture clear to everyone affected.