Meta's 7M-Selling Smart Glasses Drop Camera to Save AI Devices

Removing the camera from a leading AI wearable changes the training and inference pipeline for ambient multimodal models. Meta's 7 million-unit installed base gave it a vision data moat; a camera-free tier signals real constraints on when and where always-on AI can be deployed.

Meta Platforms is expected to use its annual Connect conference on Wednesday, 2026-09-23, to reposition its AI wearable strategy around privacy by unveiling camera-free smart glasses in two styles, reportedly named Clubmaster and Burbank. The move arrives as the company tries to hold its early lead in AI devices while answering a wave of legal, regulatory and public pushback over covert recording. With more than 7 million units sold last year through its EssilorLuxottica partnership, Meta has the largest installed base in a market where its main competition comes from Chinese firms RayNeo and Xiaomi. The camera-free models are not the only expected hardware. Analysts also anticipate a refreshed Meta Ray-Ban line retaining cameras and an ultra-light augmented-reality pair code-named Phoenix. Days earlier, Snap detailed new partnerships and software for its US$2,195 Specs AR glasses, reinforcing that the competitive window for ambient AI devices is narrowing. Meta's camera-free bet is therefore less about abandoning visual capture than about segmenting the product line for regulated spaces, cautious consumers and markets where recording anxiety is strongest.

“

Days earlier, Snap detailed new partnerships and software for its US$2,195 Specs AR glasses, reinforcing that the competitive window for ambient AI devices is narrowing.

The privacy challenge is concrete. U.K. restaurants, pubs and theatres have imposed restrictions, and a proposed California class action alleges Meta misled users about privacy protections and that human annotators training its AI viewed and labeled private moments, including changing clothes, after recordings were uploaded. Meta contests the allegations and plans to defend itself. Importantly, analysts question whether removing the camera solves the privacy problem at all; the devices still collect microphones and audio, which remain a surveillance vector and can be used without visual consent.

The commercial tradeoff is significant. Photo and video capture rank among the most popular use cases, so a camera-free SKU may blunt demand even as it reduces legal friction. At the same time, retaining audio collection means the new product could satisfy only part of the privacy critique. If the camera was the primary flashpoint for bans and lawsuits, removing it could lower friction in venues and give Meta a defensive argument, but it will not fully neutralize concerns around AI recording, data annotation and ambient data ingestion.

For AI systems, the sensor architecture of consumer wearables determines what training and inference data exist. A camera-free model may appeal to users who want hands-free audio assistance, navigation or voice control without being seen as recording. But it exits the multimodal visual pipeline that underpins the most advanced contextual AI features. Meta's 7 million-plus installed base is a strategic asset because it generates multimodal telemetry and feedback; diluting that with camera-free devices alters future model development.

The timing also reflects competitive pressure from Snap's Specs and Chinese rivals, which continue to ship camera-equipped alternatives. Meta appears to be giving consumers a choice between privacy and capability, while preserving its flagship visual AI line through Ray-Ban updates and Phoenix. That dual-track strategy could help Meta defend market share across regions with differing privacy norms, especially Europe and the U.K.