I stared at a wall.
A smiling molecule popped out of a ceramic vase, floating in the air right in front of me, complete with blinking googly eyes.
It started explaining ceramic material science.
The whole thing was generated by Gemini AI, rendered in 3D, and projected through glasses plugged into a phone-sized brick.
Project Aura.
This setup is only one piece of Google’s expanding AR puzzle.
Coming this fall, alongside the high-profile Aura unit, are wireless AI frames co-designed with Samsung, Warby Parker, and Gentle Monster.
They’re gunning straight at Meta’s Ray-Bans.
I spent days wearing prototypes at Google’s Mountain View HQ.
The December demo was cool, but this latest dive felt different.
Sharper.
Maybe even ready to redefine the category.
Will this make us want to wear them?
There’s a fear around smart glasses.
Cameras in faces. AI watching everything. It feels threatening.
Does Google have a plan for that discomfort?
The I/O demos suggested yes, or at least, a better attempt.
These might actually become daily drivers.
Smart glasses are flooding the market in 2026, mostly ignored or feared.
But the Gemini integration changes the stakes.
It’s not just the hardware, it’s the software foundation—stuff Apple is reportedly eyeing too.
I talked to Shahram Izadi, head of Google XR, and Jay Kim from Samsung’s MX division.
They sound confident.
AI smarts beat Meta’s tricks
I wore the prototype “Intelligent Eyewear.”
Names for the consumer models? TBD.
Warby Parker and the others will assign those this fall.
Like Meta’s gear, they have cameras, mics, and hidden speakers.
But there’s a split here.
Some versions have screens. Some don’t.
The real winner isn’t just voice commands or plant identification (we’ve seen that), but the bridge to apps.
Google Keep. Google Calendar.
Your phone data appears in the glasses, or on a watch.
Context carries over.
A translation demo was slick.
I spoke to people nearby, and the system flagged French and Portuguese automatically.
Captions hit the screen in the display version.
Audio translations followed, matched to the speaker’s gender and cadence, with just enough delay to keep the conversation real.
AI widgets float nearby too, scrollable stacks showing weather or quick notes.
Lighter than expected
Weight matters.
Meta’s Ray-Ban Displays feel heavy.
These? 49 grams.
Close to the standard non-display Ray-Bans, which is great for all-day wear.
The prototypes I tried had a single micro-display.
Surprisingly comfortable, even smaller than Meta’s bulkier options.
The actual designs look promising.
Warby Parker went with rounded, thicker frames that look normal enough to wear out.
Gentle Monster chose an oval shape, bolder, distinct.
The prototype on my face? Generic black. Functional. Ray-Ban Meta clones in silhouette, but better specs under the hood.
Phones, watches, and photos
Snap a pic with the glasses.
It pops up seconds later on a nearby Pixel Watch.
The watch acts as a viewfinder, which solves the blind-spot problem of taking photos without a screen.
I tried a Nano Banana effect—suddenly 1930s cartoons were eating sushi in my hallway.
The processed image hit the phone and the glasses in under a minute.
Controls live on a touchpad pad on the temple arm.
Swipes, holds, double taps.
Volume up/down with finger slides.
No neural band, like Meta offers.
That’s a gap, though the watch controls help fill it.
Dual displays are coming… maybe
There are also dual-display prototypes.
Heavier. Clunkier.
But the resolution is nice.
I watched a Spider-Man clip rendered in 3D by YouTube.
Tilting my head shifted the 3D model in my hand.
Real AR spatial interaction, not just a screen on your face.
No launch date yet.
Izadi hints it could be next year.
Or the one after.
Google likely wants to sell a range: no-display, single-display, dual-display.
Tiered pricing, different use cases.
What about privacy?
This is the elephant in the room.
Wearables raise flags. Cameras do worse.
Meta has had plenty of PR stumbles here.
Izadi admits privacy requires “user education” and designing protection from day one.
They’re adding an LED light for bystanders so people know when they’re being filmed.
AI fraud detection is on board too.
Rules for the glasses will mirror Gemini on phones, mostly.
Always-on modes make things trickier, obviously.
Kim from Samsung notes, “We need to raise the bar.”
The expectation on Google is high because, well, it’s Google.
Prescription lenses should be easy, thankfully.
I wore my custom myopic inserts.
Fit was solid.
Google leans on the Warby Parker partnership to handle this, proving displays + Rx glasses work.
Meta struggles here; Ray-Ban Display prescription options are limited.
Google might finally fix this pain point.
Health features?
Google Health ties into this.
Fitbit data, sleep metrics, coaching.
Right now? No immediate integration into the frames, but it’s on the roadmap.
Kim envisions a web connecting rings, watches, phones, and glasses.
“Everything comes together,” he said, “without overwhelming you.”
Meta pairs with Garmin.
Samsung will announce more designs in July at Unpacked.
Health will come eventually.
Project Aura changes the game
Let’s go back to the floating molecule.
That was Project Aura, made with Xreal.
It’s fascinating.
Imagine a VR headset that shrinks into AR glasses powered by a separate Android brick.
Not a phone.
A dedicated Android XR computer puck.
It replaces the tethers.
No cables.
You connect the glasses to this puck.
It runs Android apps, supports WebXR.
It does what Meta’s Project Orion aimed for, but actually available this year.
Hand tracking replaces neural bands.
You pinch and scroll in the air.
Field of view? 70 degrees on microOLED screens.
Wider than Xreal or TCL.
You forget the edges are there.
You can play Steam Deck games on it.
Hollow Knight on the “big screen,” a YouTube guide floating next to it.
Five apps open at once.
Gemini watches your gameplay and gives advice.
Plug in an iPhone.
Analyze files with Gemini while browsing.
It’s a pocket computer that thinks about what you see.
Vibe coding in VR
The craziest demo was coding in VR.
Using Gemini Canvas and WebXR.
A guy created a 3D drawing app in Chrome in minutes.
Vibe coding.
Sketching lines in air became a musical instrument app, Gemini Melody.
The molecule app was built in a week.
It showed how fast we can now build spatial experiences.
If coding in 3D space is this easy, developers are going to go crazy.
Aura launches this year. Price unknown. Likely expensive.
But it proves the form factor works.
Izadi calls moving the puck logic directly into the phone frame an “obvious next step.”
Kim says the phone stays central for years.
I agree.
The connection between these devices is just warming up.
The wireless frames come soon.
The Aura unit arrives later.
The landscape is shifting fast.
“It has to be very natural.”
Jay Kim, Samsung MX Head
The real test isn’t specs.
It’s whether we’ll actually wear them without feeling weird.
Google has a strong chance this time.
But will the average person care enough about 3D molecules or AI captions?
Maybe.
Or maybe not yet.
