On July 26, Bloomberg's Mark Gurman reported that Apple has delayed its first smart glasses (codenamed N50) to WWDC June 2027, with sales in late 2027 — a roughly six-month slip. The sole reason is privacy. Apple is prototyping on-device processing for all visual data, no server uploads, no use of recordings to train AI models, and a possible camera-free variant. The delay is a direct response to Meta's Ray-Ban privacy scandals. DrafterDaily explains what Apple is building, how it compares to Meta's approach, and what the delay means for the AI glasses market.
On July 26, Bloomberg's Mark Gurman reported that Apple has pushed the reveal of its first smart glasses — codenamed N50 — from a planned late-2026 event to WWDC June 2027, with general availability following in late 2027. That is roughly a six-month slip. The reason, according to multiple sources, is privacy: Apple's Vision hardware group considers it the top design priority, and the features the team is prototyping are not ready.
The framing in tech media has been 'Apple delays smart glasses.' A more accurate framing is 'Apple made the most consequential product philosophy bet in the AI glasses market.' The two frames are the same event and very different stories.
The N50's timeline revision is a reveal delay, not a product cancellation. Apple's hardware roadmap for the glasses remains intact; the company is still building them and still plans to launch them. What changed is the engineering prioritization: features that were on the roadmap for post-launch updates are being moved into the launch product, because the team has concluded that shipping privacy-incomplete is a worse outcome than shipping six months later with privacy solved.
The specific features Apple is prototyping: on-device processing for all visual data captured by the glasses' cameras, with no uploads to Apple servers in standard use; a policy of not using customer recordings to train Apple's AI models without explicit opt-in; a tamper-responsive camera indicator (a hardware mechanism, not just a software flag, that detects and logs if the recording light has been disabled or covered); and a possible variant of the glasses that uses environmental sensors without a traditional camera lens for users who want location and context awareness without video capture capability.
Meta's Ray-Ban smart glasses have been commercially successful by the metrics Meta reports — they are the market share leader in AI glasses as of mid-2026. They have also accumulated a persistent record of privacy criticism: concerns about undisclosed recording in public spaces, questions about how video captured by the glasses is processed and stored, and a camera indicator system that researchers demonstrated could be tampered with in ways that made the recording light invisible without disabling the camera.
On July 7, 2026, Meta issued a software update specifically disabling the camera if the recording indicator light is detected as tampered with or obstructed. The update was a response to the demonstration, not to any reported misuse. But the fact that the update was necessary is the data point Apple's hardware teams are working from. The question Apple is trying to answer: what would a smart glasses product look like if the privacy architecture was designed first, rather than patched after shipping?
Meta's 12–18 month first-mover advantage in the AI glasses market is real. They have existing customers, existing partnerships, existing developer relationships, and a product that people have already decided to wear (or decided not to wear) in public. That is a durable head start.
The question is whether the AI glasses market is won by first mover or by first product that people trust enough to wear all day, everywhere. Meta's glasses are on the faces of a specific demographic of early adopters. A large number of potential customers have been watching the privacy controversy and making a different decision: wait. Apple is betting that 'wait' group is large, that those customers will buy when a product they trust is available, and that being second with a better privacy story is more valuable than being first with an incomplete one.
For enterprise buyers evaluating wearable AI — where employees and customers are going to encounter the technology in professional settings — Apple's approach may be the more credible one regardless of what individual consumers decide. A CISO evaluating whether to allow smart glasses in a corporate environment will find it easier to say yes to a product with on-device processing, no server uploads, and a hardware-verified camera indicator than to one that required a post-launch software patch to address a basic trust issue.
Timeline for enterprise and consumer planning: Apple N50 reveal at WWDC June 2027; general availability late 2027. If you're evaluating smart glasses procurement for a corporate rollout, you have 18 months before Apple's product is available to evaluate. If you're making personal purchase decisions, Meta's Ray-Bans and the competing Google and Samsung products are the options available now; the Apple product is a 2027 decision.
The AI glasses market will not be decided in 2026. The companies that are in the market now will build customer relationships, accumulate usage data, and refine their products. Apple's six-month delay is a bet that 2027 is early enough to win a market that is still forming. Whether that bet is right depends on how fast the market grows in the meantime and whether Apple's privacy-first product story resonates with the customers who are waiting to see what Apple does. The delay is, in effect, the product decision.
Based on Mark Gurman's July 26, 2026 Bloomberg report, Apple plans to reveal the N50 smart glasses at WWDC June 2027, with general availability in late 2027. These dates are based on Apple's current internal timeline as of late July 2026; Apple has not made a public announcement. Product timelines can shift further before official announcement, and Apple does not comment on unreleased products.
What the biggest product decisions mean before they ship — explained clearly.
CVE-2026-0768 is an unauthenticated root RCE in Langflow. It was disclosed in January, the fix has shipped through seven releases, and attackers are hitting it in September — because low-code AI middleware became critical infrastructure without acquiring a patch owner.
Opposition to data centres is not a referendum on AI. It is a cost-allocation dispute — and PJM's capacity auction is the mechanism that turned an abstract argument into a number on 67 million households' bills.
OpenAI's first published benchmarks for its custom inference chip are real, from a public benchmark, and normalised on nameplate TDP rather than measured power. OpenAI disclosed that itself, in a sentence nobody is quoting.