Meta and EssilorLuxottica plan to release AI-powered smart glasses under the Oakley brand, priced at around $360, CNBC reported last year. The higher price point reflects added weather resistance suited to Oakley's athletic buyer base, a step up from the Ray-Ban entry point of $299. A Prada collaboration is also planned through the same EssilorLuxottica licensing structure, though no announcement timeline has been confirmed publicly.

At the premium end, Meta was reportedly developing a higher-end Ray-Ban Meta variant priced around $1,000, The Verge reported last year. The full range from $299 to near four figures, spanning performance, fashion, and lifestyle frames, reflects a deliberate push to cover multiple buyer segments under one hardware ecosystem rather than occupy a single style niche.

The EssilorLuxottica partnership dates to 2019, when Meta first teamed with the company to build the Ray-Ban Meta line, per CNBC. The Oakley and Prada extensions run through that same relationship. Whether the software stack and data practices carry across those new brand lines has not been confirmed independently, but given the shared underlying partnership, privacy advocates and regulators are treating the expansion as a scale-up of the same system, not a fresh start.

More frames and more price points also mean a larger user base feeding data into that system. That's the context the Texas investigation lands in.

The AI features don't run on-device. Any time a user invokes Meta AI, say with a "Hey Meta" voice command, that footage is sent to Meta's servers for processing, the EFF noted earlier this year. There's no local processing option. Media captured through the glasses is also auto-imported to the Meta AI mobile app by default, meaning most users are pushing content to the cloud without making an active choice to do so.

In April 2025, Meta updated its policy for existing Ray-Ban Meta owners on two fronts. The AI camera mode is now always active unless users manually disable the "Hey Meta" voice function. The option to prevent voice recordings from being stored in the cloud has been removed entirely, The Verge reported last year. Voice transcripts and audio recordings are retained for up to one year by default; users can delete them manually, but cannot opt out of storage.

The LED recording indicator designed to signal active capture to bystanders does not illuminate during always-enabled mode and can be physically obscured, the National Law Review noted this week, citing the AG's announcement. Neither wearers nor the people around them have a reliable visual signal when the AI camera is processing.

Some footage used for AI training goes through human review. An investigation by Swedish journalists found that workers at Meta's Kenyan subcontractor Sama reviewed and annotated user-captured video, including footage of intimate moments, the EFF reported earlier this year. Meta told the BBC this is consistent with its terms of service. Spokesperson Albert Aydin told The Verge that photos and videos on a user's camera roll are not used for training unless shared to another product, but media stored via the Meta AI app falls under that product's policies, not a blanket exemption. Footage saved to a phone's camera roll may also sync separately to Apple or Google cloud services depending on the user's settings, adding another potential access point, per the EFF.

Users do have some controls. The Meta AI app includes a privacy settings page where "Cloud media" can be disabled, stopping photos and videos from being sent to Meta's cloud. Disabling the "Hey Meta" voice command deactivates always-enabled AI camera mode. What users cannot do is opt out of voice recording storage, or guarantee that footage used for AI training won't go through human review. Those limits apply to current owners now and would apply to anyone buying Oakley or Prada variants if those ship under the same framework.

The Texas AG's Civil Investigative Demand targets whether the glasses' always-on recording practices, facial geometry collection, and privacy representations violate Texas consumer protection law, the National Law Review reported this week. The scope of the demand and Meta's response have not been publicly disclosed.

This is the second major enforcement action the Texas AG's office has directed at Meta over biometric data practices in less than two years. The first ended in a $1.4 billion settlement in July 2024, then the largest privacy recovery in Texas history, over Meta's unlawful capture of facial recognition data through Facebook, per the National Law Review. The AG's office has framed the current glasses investigation as a continuation of that same regulatory concern, not a new category of risk.

The sharpest issue in the probe is a reported internal feature called "Name Tag," which would enable the glasses to identify individuals within the camera's field of view, including bystanders with no relationship to Meta and no opportunity to consent, according to the National Law Review, citing New York Times reporting on internal Meta plans. No official timeline for the feature has been confirmed, and Meta has not made any public statement on it. The EFF warned earlier this year that adding facial recognition to smart glasses would "obliterate the privacy of everyone." That warning now reads as a reasonable preview of where regulators landed.

Texas is the active case. It may not stay the only one. Illinois, Texas, and Washington have standalone biometric privacy laws, and California, Illinois, and Washington give consumers a direct right to sue for violations, per the National Law Review. Whether other states follow will depend partly on what the Texas CID surfaces.

The Texas investigation is early-stage. But it carries specific weight: it arrives backed by a $1.4 billion precedent, issued by an office that has already demonstrated willingness to pursue substantial monetary relief over biometric collection without proper notice and consent.

Three things will determine whether Meta's smart glasses platform scales as planned or runs into structural friction: whether the always-enabled defaults survive regulatory scrutiny, whether "Name Tag" advances toward launch and how enforcement responds, and whether the gap between Meta's "designed for privacy" marketing and its operational data practices becomes a consumer trust problem at scale.

The Texas AG's office is now testing that marketing claim directly, per the National Law Review. A public disclosure from the CID process, a confirmed Prada launch timeline, or any movement on "Name Tag" would each shift the story considerably. Until then, the Oakley and Prada frames are coming, the privacy architecture behind them is already under investigation, and the question for prospective buyers is the same one regulators are asking: what actually happens after the camera opens.