Large Platforms Absorb Ambient AI
Sandbar
This market is being decided by distribution, not product polish. Once Amazon and Meta started buying ambient AI startups, the contest shifted from who can build a better recorder to who can bundle memory, voice capture, and assistant workflows into hardware people already wear, and into software rails they already use. That is a much harder game for an independent device company with limited marketing budget and no installed base.
-
Amazon bought Bee in July 2025, then started adding features that turn spoken moments into email and calendar actions. Bee began as a $49.99 wearable with a subscription, but inside Amazon it gains Alexa, Echo, and Prime level distribution that can crush customer acquisition costs for any startup competing for the same use case.
-
Meta bought Limitless in December 2025 and stopped new hardware sales, folding the team into Reality Labs. That matters less because Limitless was huge on its own, and more because it shows memory capture is being treated as a feature for glasses and other Meta wearables, not as a standalone pendant business.
-
The contrast with Plaud is revealing. Plaud has already scaled to an estimated $250M annualized revenue by September 2025 by serving offline professionals like plumbers, lawyers, and doctors, while Limitless was at an estimated $2M in revenue in April 2025. The surviving independents are the ones tied to a narrow, urgent workflow, not broad ambient consumer AI.
From here, the category is likely to split in two. Large platforms will absorb the broad consumer memory assistant into glasses, earbuds, phones, and home assistants, while independents that survive will look more like workflow tools with hardware attached. For Sandbar, the path forward is to own a specific repeated behavior so tightly that it cannot be reduced to just another platform feature.