Meta CEO Mark Zuckerberg introduced the company’s latest AI-powered smart glasses at the Connect 2025 keynote, but the live demos didn’t quite go as planned.
TLDR:
- Meta launched its new Meta Ray-Ban Display glasses, priced at $799, with high-resolution displays and AI features.
- Live demos failed on stage, including an AI assistant misfiring and a missed video call from the CTO.
- The mishaps were due to a self-inflicted server overload, not poor Wi-Fi, as initially claimed.
- Meta CTO Andrew Bosworth clarified the root causes, saying they were demo-related, not product faults.
What Happened?
At Meta’s Connect 2025 event in California, Zuckerberg unveiled the Meta Ray-Ban Display, calling it the company’s “first AI glasses with high resolution.” But two high-profile demo failures overshadowed the moment, drawing criticism and sparking confusion.
Despite a packed audience and a live stream with 4,000 viewers, the smart glasses’ key features failed during crucial moments. Zuckerberg was left fumbling on stage when the glasses couldn’t answer a WhatsApp video call. Meanwhile, another demo involving a cooking assistant AI glitched, skipping recipe steps and failing to respond properly.
Meta’s Big Bet on AI Glasses
Meta is pushing further into the wearables market with its newest Ray-Ban smart glasses. The Meta Ray-Ban Display model comes with a 5,000-nit brightness screen, a hands-free interface powered by the Meta Neural Band, and integrated AI features that can help with tasks like messaging, music control, and even cooking instructions.
The event also introduced two other devices:
- A sports-oriented pair of Oakley-style glasses called Meta Vanguard.
- An upgraded version of the existing Ray-Ban Meta glasses, now priced at $379.
Zuckerberg’s presentation started with confidence. He showed off features like live texting, emoji responses, and gesture-based controls, claiming he could write 30 words per minute using just wrist movements.
But then, the show went sideways.
What Went Wrong on Stage?
One of the most talked-about moments came when Zuckerberg failed to answer a WhatsApp call from Meta CTO Andrew Bosworth using the Neural Band. Moments later, a cooking content creator tried to use the glasses’ Live AI to get sauce-making instructions. Instead, the assistant skipped steps, leaving the presenter confused. He even blamed the Wi-Fi.
Later, Bosworth set the record straight.
In an Instagram AMA, Bosworth explained that when the chef said “Hey Meta, start Live AI,” it accidentally triggered the same command on every pair of smart glasses in the building. Since many attendees were wearing them, this created a massive surge of traffic to Meta’s internal development server.
“We DDoS’d ourselves, basically,” Bosworth admitted. He also revealed that a rare bug caused the glasses to go to sleep at the exact moment a call notification came in, preventing the device from responding.
He emphasized that these were demo failures, not product flaws, and said the issues had been resolved since.
What TechKV Thinks?
Honestly, I kind of respect that Meta went live and unscripted. Most companies play it safe with pre-recorded demos, but Meta took the risk and paid the price. Still, it shows a level of transparency that’s rare in the tech world.
The launch had potential to be a landmark moment for AI wearables, but the glitches exposed the complexity behind “smart” products that rely on live servers and voice commands. It’s a reminder that even tech giants can trip over their own wires. If Meta wants people to believe in AI glasses, they’ll need to prove it works under pressure, not just in rehearsals.
