Google announced two Gemini AI glasses at I/O 2026: audio-only launching this fall and a display prototype. Here's what's real and what's still coming.
Google Is Back in the Glasses Game — And This Time It’s Different #
Google Glass was a cautionary tale. Expensive, awkward, and socially strange, it became the symbol of tech that moved too fast without asking whether anyone actually wanted it.
The Gemini AI glasses announced at Google I/O 2025 are a different bet. Google has learned from that era — and from watching Ray-Ban Meta frames quietly build a real user base. This time, the company announced two distinct products aimed at different timelines: audio-only glasses launching in the near term, and a display-equipped prototype that’s still in development.
This article breaks down exactly what each version is, what hardware and AI capabilities they include, when you can actually expect to own one, and what the practical difference between audio and display really means for daily use.
What Google Actually Announced #
At Google I/O 2025, Google didn’t just tease a concept video. They showed hardware, named fashion partners, and gave a rough shipping window for at least one version.
Here’s the short version:
Android XR Glasses (audio-only)— Real glasses with Gemini built in, partnered with Warby Parker and Gentle Monster. These have a camera, microphone, speakers, and Gemini access via voice. Launching in 2025.Display prototype— A version with an integrated heads-up display (HUD) that overlays information in your field of view. Shown on stage, but not close to commercial release.
- ✕a coding agent
- ✕no-code
- ✕vibe coding
- ✕a faster Cursor
The one that tells the coding agents what to build.
These are two separate products at two very different stages of readiness. The audio version is essentially ready to ship. The display version is still a research and development project being shown to signal where Google is headed.
The Audio-Only Glasses: What You Actually Get #
Hardware and Design
The audio Gemini glasses look like regular eyewear. That’s intentional. Google partnered with Warby Parker (a mainstream eyewear brand with wide retail reach) and Gentle Monster (a high-fashion Korean brand) specifically to make glasses people would want to wear in public.
Built-in components include: Open-ear speakers— Audio plays without covering your ears, similar to how bone-conduction headphones work in concept** Built-in microphone array**— Picks up voice commands and ambient audio** Camera**— Positioned to capture what you’re looking at; feeds visual context to Gemini** Battery**— Charged via a case, like wireless earbuds
The glasses run on the Android XR platform, which is Google’s operating system layer for extended reality hardware. Android XR handles the connection to your phone and the pipeline that routes audio and visual data to Gemini.
What Gemini Does in These Glasses
The core experience is Gemini as a persistent, context-aware assistant. Because the camera captures what’s in front of you and the microphone hears what’s around you, Gemini can answer questions about your environment in real time.
Practical examples:
- You’re at a restaurant reading a menu in a foreign language — you ask Gemini to translate it
- You’re in a hardware store looking at a bolt — you ask what size it is or whether it matches something you already have at home
- You get a text while walking — Gemini reads it to you and you dictate a reply
- You’re looking at a building and want historical context — Gemini tells you about it based on what the camera sees
This is Project Astra-style functionality made wearable. Google’s Project Astra research demonstrated multimodal, real-time reasoning across video, audio, and text. The audio glasses are the first consumer product built on that foundation.
When Are They Shipping?
Google confirmed the audio glasses are targeting a 2025 launch, with fall 2025 as the target window. Warby Parker confirmed they’re the U.S. retail partner, which means they’ll likely be available in stores and online through a familiar purchasing experience.
Pricing hasn’t been officially confirmed, but given the Ray-Ban Meta glasses sit around $299, Gemini audio glasses in a similar range would be a reasonable expectation.
The Display Version: What Google Showed and What It Means #
What the Prototype Actually Does
The display prototype integrates a small transparent display into the lens — a heads-up display that overlays information directly in your line of sight. Think turn-by-turn navigation that floats in front of you, or a translation of a sign you’re looking at appearing overlaid on that sign.
On stage at I/O, Google demonstrated:
- Real-time translation of spoken words appearing as text in the lens
- Navigation directions overlaid without needing to glance at a phone
- Gemini responses shown as text in the field of view
This is closer to what people imagined when Google Glass was announced in 2012. The difference is that the underlying AI — Gemini — is dramatically more capable than anything that existed then, and the hardware miniaturization has improved considerably.
Why It’s Not Shipping Yet
Display integration is genuinely hard. Fitting a transparent optical display, the projection hardware, battery, and compute into something that weighs and looks like normal glasses requires component sizes and power efficiency that aren’t quite there for consumer pricing.
Google is being transparent about this. The display glasses were shown as a prototype and research direction, not an imminent product. The goal appears to be demonstrating the vision for where Android XR glasses head in the medium term — probably 2026 or later for anything approaching mass availability.
The Fashion Partner Strategy
One underrated detail: Google isn’t building these glasses themselves the way Apple builds hardware end-to-end. The partnership with Warby Parker and Gentle Monster is structural — these brands design the frames, control the aesthetic, and sell through their channels.
This is Google’s admission that the hardware aesthetics problem with Glass was real. They’re letting design-focused brands own that part. Google focuses on the AI layer, Android XR, and the Gemini integration.
How Gemini Makes These Different from Other Smart Glasses #
It’s Not Just a Voice Assistant
Smart glasses with voice assistants aren’t new. Amazon Echo Frames have existed for years. What makes Gemini different is multimodal reasoning — the ability to process visual, audio, and contextual information simultaneously and respond coherently.
Previous smart glasses could tell you weather or play music. Gemini can look at what’s in front of you and reason about it. That’s a different category of capability.
Live Context vs. Queued Requests
Most AI assistants work on a request-response model. You say a command, you wait, you get an answer.
Gemini in the glasses is designed for live context — meaning it can maintain awareness of an ongoing situation, not just answer one-off questions. If you’re in a meeting, it can track what’s being discussed and provide relevant follow-up context. If you’re walking through a store, it can maintain a thread of what you’re shopping for.
This is what Project Astra was built toward, and the glasses hardware is the natural form factor for it.
Phone Dependency vs. On-Device Processing
The audio glasses require a connected Android phone. Processing happens partly on-device and partly via the Gemini cloud infrastructure. This means:
- You need your phone with you
- The experience depends on connectivity
- Battery constraints are partially offloaded to the phone
The display version will likely move more processing on-device over time, but for the audio version, the phone is a required companion.
Audio vs. Display: A Direct Comparison #
Here’s what actually differs between the two versions:
| Feature | Audio Glasses | Display Glasses |
|---|---|---|
| Availability | Fall 2025 | Prototype / TBD |
| Visual output | None (audio only) | Heads-up display overlay |
| Camera | Yes | Yes |
| Microphone | Yes | Yes |
| Speakers | Yes | Yes |
| AI model | Gemini via Android XR | Gemini via Android XR |
| Retail partners | Warby Parker, Gentle Monster | Not announced |
| Price | ~$299 (estimated) | Not announced |
| Form factor | Normal glasses | Still optimizing |
Built like a system. Not vibe-coded.
Remy manages the project — every layer architected, not stitched together at the last second.
The audio glasses are the practical buy for 2025. The display version is the more futuristic option that most people will wait years to actually use.
What This Means for Developers and AI Builders #
The Gemini API Is the Real Foundation
Both versions of the glasses run Gemini. That matters to anyone building on top of Google’s AI infrastructure, because the same model powering these glasses is available via the Gemini API — and accessible through platforms that make it much easier to integrate.
Developers can build applications that tap Gemini’s multimodal capabilities without needing to manage hardware. The glasses are one surface; there are many others.
Building With Gemini Without Managing Infrastructure
If you’re interested in building AI agents or workflows that use Gemini, MindStudio is worth knowing about. It’s a no-code platform that gives you access to Gemini and 200+ other AI models — Claude, GPT-4o, and others — without needing separate API keys or accounts. You can build agents that use Gemini’s vision and reasoning capabilities and connect them to business tools like Google Workspace, Slack, or HubSpot — all through a visual builder. The average build takes under an hour.
This is relevant in the context of Gemini glasses because the assistant intelligence in those glasses is the same Gemini you can build with today. Understanding what Gemini can do in a glasses context gives you a clearer picture of what it can do in an agent context.
You can start building free at mindstudio.ai.
Frequently Asked Questions #
What are Google Gemini AI glasses?
Google Gemini AI glasses are smart eyewear that integrate Google’s Gemini AI model for real-time audio and visual assistance. They come in two versions: an audio-only version launching in fall 2025 (through partners Warby Parker and Gentle Monster), and a display prototype that overlays information directly in the lens. Both run on Android XR and use Gemini to answer questions about what you’re seeing and hearing.
When will Google Gemini glasses be available to buy?
The audio-only version is targeting a fall 2025 launch in the United States, through Warby Parker’s retail channels. The display version has no confirmed release date and is currently in prototype stages. Google presented the display version at I/O 2025 as a product direction, not an imminent launch.
What’s the difference between the audio and display versions?
The audio glasses have no screen — they deliver all information through built-in speakers. Gemini responds verbally to your questions and can describe what the camera sees. The display version adds a transparent heads-up display that overlays text and visuals in your field of view — showing things like translated text, navigation directions, or AI responses as readable overlays. The display version is significantly more complex to manufacture and is not yet available commercially.
Do Google Gemini glasses require a phone?
Yes, the audio glasses require a connected Android phone. They use Android XR as the operating system layer, and processing is split between the glasses hardware and your phone, with some inference happening in Gemini’s cloud infrastructure. You’ll need a compatible Android device and reliable connectivity for the full experience.
How do Google Gemini glasses compare to Ray-Ban Meta glasses?
Both are audio-first smart glasses with cameras and built-in speakers. Ray-Ban Meta glasses use Meta AI as the assistant; Google’s glasses use Gemini. The key difference is Gemini’s multimodal reasoning capability — it can process what the camera sees and respond with contextual, visual understanding in a more sophisticated way than current Meta AI glasses. The design approach is similar: fashionable frames made with established eyewear partners. Ray-Ban Meta launched in 2023 and has already gone through two generations; Google’s audio glasses are entering a more established market category.
Is the display version of Google’s AI glasses the same as Google Glass?
Not really. Google Glass (2013–2015) ran basic apps and had a fixed prism display above one eye. The new prototype display glasses are built around Gemini — a large multimodal AI — and use a more integrated display design. The fundamental concept of overlaying digital information onto real-world vision is similar, but the AI capability, hardware design, and market strategy are substantially different. Google also learned from Glass’s social and commercial failure, which is why they’re moving more cautiously with the display version.
Key Takeaways #
- Google announced two Gemini glasses at I/O 2025: audio-only (shipping fall 2025) and a display prototype (no release date).
- The audio version looks like normal glasses — built by fashion partners Warby Parker and Gentle Monster — with a camera, microphone, speakers, and Gemini access via voice.
- The display version overlays information directly in your field of view but remains a prototype; don’t expect to buy one in 2025.
- Both versions run Android XR and use Gemini’s multimodal capabilities to understand what you’re seeing and hearing in real time.
- The same Gemini model powering these glasses is available to developers and builders via the Gemini API — and accessible through platforms like MindStudiowithout managing infrastructure. - If you want to experiment with building Gemini-powered applications or agents today, MindStudio lets you start for free with no API key required.