Menlo Park, California, September 18, 2025 Meta’s much-anticipated Connect 2025 keynote, headlined by CEO Mark Zuckerberg, was rocked by several high-profile technical misfires during live demonstrations of its new AI-powered Ray-Ban smart glasses. What was meant to be a showpiece unveiling turned into a cautionary reminder of the risk of live product demos.
What Went Wrong
Meta introduced a new line of wearable tech, including:
-
The Meta Ray-Ban Display glasses with an in-lens display controlled by gestures via a Neural Band wrist accessory.
-
Upgrades to its Ray-Ban Gen-2 smart glasses, plus sport-oriented versions in partnership with Oakley.
During the live demo section, two major glitches stood out:
-
Cooking/Recipe Demo Failure
When Meta teamed up with influencer Jack Mancuso to showcase a “Live AI” feature that was supposed to guide recipe preparation, the system malfunctioned. Mancuso asked what he should do first with the ingredients, but the AI incorrectly replied that the “base ingredients” had already been combined when they hadn’t. After repeated requests, the AI continued to give the wrong instruction (“grate some pear”) instead. The demo ended with Mancuso apologizing, joking that the WiFi may have been messed up, and passing things back to Zuckerberg. -
Video Call Demo Breakdown
Later in the presentation, Zuckerberg tried to demonstrate the glasses + Neural Band’s capability to receive a WhatsApp video call via gesture control. While text messaging worked, the video call portion repeatedly failed to connect. Zuckerberg tried multiple times but the glasses did not show the incoming call appropriately. The audience heard notification rings, but the UI never responded. Andrew Bosworth, Meta’s CTO, later described the cause as “WiFi issues” and the demo as “a demo fail, not a product failure.”
What Might Have Caused the Failures
A few probable causes emerged from Meta’s own explanations, expert commentary, and post-event analysis:
-
Network Overload / WiFi Problems: Meta said that during the cooking demo, when the AI activation was triggered, it was inadvertently activated on all glasses in the building creating a kind of self-inflicted “DDoS-like” overload. The network environment during the live event was far more crowded than during rehearsals.
-
Bug in the Operating or Display Logic: For the video call demo, a rare bug caused the glasses’ display to “sleep” or fail to wake in time when the notification arrived. The UI didn’t respond even though the system had gotten the notification.
-
Live Demo Risk in General: Demonstrations in front of a live audience with many variables (network traffic, hardware variance, environmental interference) increase the chance of something going wrong even if rehearsals go well. Some commentators noted that devices like smart glasses, when working off wireless connectivity and external guidance (AI, cloud services), are especially vulnerable.
Reactions & Consequences
-
Public & Media Response: The glitches immediately caught the attention of tech media, social platforms, and analysts, many pointing out the irony of “smart” glasses failing during a smart demo. Memes and comparisons to other famous tech demo fails were swift.
-
Meta’s Acknowledgement: Meta’s CTO Bosworth later emphasized that the failures did not reflect fundamental flaws in the product, that the devices work under test conditions, and that fixes are underway. Zuckerberg himself joked about the WiFi, admitted the embarrassment, and pressed on with the presentational flow.
-
Sales & Product Outlook: Despite the hiccups, Meta proceeded with its product reveals. The Ray-Ban Display glasses are priced at $799, with a launch date set for September 30 in the U.S. Updated Ray-Ban smart glasses, and Oakley-collaboration models, were also announced. Analysts remain mixed: optimistic about potential, but cautious about user adoption, price sensitivity, and how real-world performance will hold up.
Key Takeaways
-
Even well-funded companies with resources for rehearsals can be caught off-guard in live settings, especially with connected AI/IoT devices.
-
Network environment control is critical for live demos; reliability under stress (many devices, many users) is a major challenge.
-
Public perception and confidence depend not just on product specs, but on how smooth the product looks in real use. High expectations risk high scrutiny.
Implications for Meta & Beyond
Meta has positioned these smart glasses as part of its vision of wearable AI, “superintelligence”, and augmented reality being more integrated into everyday life. The demo failures do not seem to derail that vision, but they do raise questions about:
-
Whether the hardware/software is mature enough for mass adoption;
-
How robust the devices are in real-world, less controlled environments;
-
How Meta handles reliability and trust (especially when glitches are visible to the public);
-
The importance of fallback or contingency strategies in public launches.