Meta glasses are at the center of a new controversy, as the company quietly announced that it will impose a rate limit on a key feature that runs entirely locally on the device — without requiring any connection to its servers. Specifically, the Conversation Focus will be limited to just three hours per month for users who don’t pay a subscription, while those who sign up for Meta One Premium ($19.99/month) will have access to up to 15 hours per month. The move has sparked strong reactions, as many users consider it unreasonable to pay a subscription for hardware they already own.
See also: Meta: Preview of Artificial Intelligence Features in Ray-Ban Smart Glasses
Conversation Focus is a feature that uses the built-in open-ear speakers , beamforming technology , and real-time spatial processing to amplify the speaker’s voice in noisy environments. The crucial point is that this feature does not require an internet connection — it runs exclusively on the chips inside the glasses. Tests have confirmed that Conversation Focus continues to work even with Airplane Mode enabled on the smartphone. This makes Meta ’s claim of a “rate limit” highly questionable, since there is no server-side cost to justify the limitation.
Meta Glasses and the business model behind the rate limit
Meta already accounts for 76% of global smart glasses shipments in 2024, according to International Data Corp , with more than 7 million pairs sold in 2025. The company is facing significant financial pressure from its investments in artificial intelligence , which recently led to the layoffs of about 8,000 employees — 10% of its total workforce . At the same time, the company reduced the price of the new glasses by $80 by removing the Ray-Ban brand name , with the new models ranging from $299 (Fury/Adventurer) to $399 (Starfire Kylie Edition), while the premium Ray-Ban Display starts at $799 .
Analysts say Meta is attempting to turn hardware into a subscription revenue platform, a model that could impact the entire wearable AI market. The Meta One Premium is expected to expand to other features, such as detailed responses (richer visual descriptions) and phone call captioning. However, the move comes with serious risks: users on Reddit are already reporting that the feature is showing up as “not available” after purchase, calling it a “ripoff.”
See also: Meta becomes Qualcomm's first customer for Dragonfly chips
Meta Glasses, privacy and the risks for users
Beyond the financial issue, Meta glasses have also come under fire for privacy. The company quietly removed the option in 2025 that allowed users to opt out of storing voice and video data in the cloud, making it mandatory. Voice transcripts are kept for up to a yearunless the user manually deletes them. In addition, the glasses have gained the infamous reputation of “pervert glasses” due to incidents of secretly recording people without their consent — although the maximum recording duration per session is limited to 3 minutes.
Of particular concern is the issue of facial recognition: Meta had secretly embedded facial recognition code in millions of smartphones, which it later quietly removed. Experts warn that if this feature is reinstated, it could replicate on a massive scale the so-called “Harvard Doxxing experiment,” in which students were able to identify strangers in public spaces using AI. Meta has not responded to questions about whether it plans to put other on-device features behind a paywall.
The competitive landscape is shaping up quickly: Google, Samsung , and Apple are developing their own smart glasses, while OpenAI is designing AI earbuds. Meta’s smart glasses already support live translation in 20+ languages, turn-by-turn navigation, and hands-free photography — features that make them attractive. But according to The Verge, imposing a rate limit on a serverless mode is a move that undermines user trust and sets a dangerous precedent for the entire wearable AI market.
See also: Lawsuit against Meta for Ray-Ban Meta glasses
In short, Meta ’s strategy with the Meta smart glasses reveals a company that is attempting to turn hardware into a platform for ongoing revenue — even if that means imposing artificial limitations on features that cost the company nothing. For users, the message is clear: buying an AI device today does not guarantee full access to its capabilities tomorrow, without an additional monthly fee.
