In recent years, the focus of Google’s annual I/O conference has shifted from Android developers to development within Google’s AI ecosystem, whether the AI runs on-device or on Google Cloud. This year’s event included updates to AI coding tools such as Antigravity and the introduction of new models, including Gemini 3.5 Flash and the Gemini Omni multi-modal models.

Google I/O 2026 also showcased some of the company’s work in smart wearables and spatial video, including AI smart glasses, the Project Aura AR glasses, and the Beam spatial video conferencing system (co-developed with HP). These developments are a good reminder of how the cutting edge in consumer and enterprise XR gear is closely tied to the latest advances not only in 3-D visuals, but even more importantly in AI functionality. I have said for years that AI and XR are a natural fit, and Google I/O reinforced that.

Google Advances AI Glasses at Google I/O

Google talked a lot about AI in glasses at Google I/O, including giving an update on the AI glasses partnerships it has built with Samsung and with eyeglasses makers Warby Parker and Gentle Monster. While Google didn’t provide any exact details on pricing or availability, it did share one design from each eyeglasses partner, which I believe is just the beginning of the styles that both Warby Parker and Gentle Monster will offer. Google calls these devices “audio glasses” to differentiate them from display glasses, which feature either a monocular or a dual electronic display on the lens.

Google demoed some interesting features of these glasses powered by Gemini, including the ability to capture photos with the glasses that you can then see on your WearOS smart watch. This strikes me as a very clever way to leverage wearables as an ecosystem of AI devices. I could also see Google implementing existing WearOS features, such as pinch detection, to make its AR and AI user interfaces feel more natural. I believe that Gemini, especially Gemini 3.5, will prove to be one of the best interfaces for AI smart glasses, empowering users to do more than they can today with competing products like Meta Ray-Ban.

Google also gave demos of its display glasses to the press and analysts. While it remains unclear whether these glasses will be launched as a Samsung product or through Gentle Monster and Warby Parker, it appears they are nearing a final design. These glasses are significantly lighter than the Meta Ray-Ban Display glasses I recently reviewed and have a much friendlier design. That said, these glasses appear to have a narrower field of view than the Meta Ray-Ban Display, but this may be a case of sacrificing FoV in favor of weight and possibly battery life. (Google hasn’t offered any figures for battery life or estimated product lifespan yet.)

Google has said that we can expect both the audio glasses and the display glasses in the fall. While pricing is not available yet, it will be important for both types of glasses to be price-competitive with Meta, considering that Meta has established this product category and consumer expectations of pricing. I believe that Google’s audio glasses will need to sell in the ballpark of $400, or ideally at $299. The display glasses can probably sell closer to $800, but that’s without a neural band — one of the best aspects of the Meta Ray-Ban Display. We have yet to see what Google will do with UI for the display glasses and how much it will leverage WearOS.

Project Aura Becomes the Tip of the Spear for Google’s AR Efforts

Project Aura is Google’s effort to be competitive in the AR space as part of its broader XR strategy, and while the company didn’t mention it from the stage, there were some major updates to the program shared at the show. I had the opportunity to try out Project Aura in January at CES, but was sworn to secrecy about it. At that time, Project Aura felt much more like a prototype with lots of room for improvement. Since then, it has become clear that Google and XREAL have worked hard to make this a much better device. Back in January, the first thing that struck me was how much smaller and lighter Project Aura was than Samsung’s Galaxy XR headset. It was also obvious that the Android XR UI was extremely consistent across Android XR devices — including Project Aura — so there was no learning curve at all.

At CES I also gave some feedback to the XREAL team that the image quality wasn’t quite good enough. I was assured that it would improve — and it has. When I tried the new version of the headset at Google I/O, the image quality was extremely sharp and crisp, and it was very easy to read text projected into the headset. While I haven’t been given detailed display specs, Project Aura’s 70-degree FoV remains among the widest in the industry. The glasses are also well-balanced, so they don’t feel too heavy on the bridge of your nose.

The compute puck, which is roughly the size of the Apple Vision Pro’s battery, has been completely overhauled. It houses the device’s compute and battery, and now the team has added XREAL’s passthrough capabilities, meaning that you can connect these glasses to any smartphone, PC, or gaming handheld as an external display. For ease of use and better enterprise security, the compute puck also now has a fingerprint sensor (particularly useful because nobody wants to type passwords into a virtual keyboard).

The thermals on the compute puck are also significantly improved thanks to a better cooling solution. In my first demo, the compute puck got noticeably hot, while the new one barely gets warm. Besides improving comfort, this is also important for limiting thermal throttling, which can affect frame rate or give users motion sickness. Speaking of frame rate, the hand tracking on Project Aura is much improved from January as well; I believe this may come in part from the major Android XR update in April, after which I noticed improved eye and hand tracking on the Galaxy XR. While these glasses don’t have eye tracking, they do have face recognition, which allows them to turn off their electrochromic dimming when you are talking to someone.

The Project Aura form factor, I believe, is going to be one that people eventually want to take with them everywhere and could even use in public. I don’t think we’ll see that in the near term, especially given the cable and the fairly large size of the glasses, but I think the long-term product vision is to wirelessly connect the headset to your phone without a cable. I believe that’s why XREAL added GPS capabilities to the compute puck, enabling developers for Project Aura to create world-scale AR experiences. Google and XREAL have a developer program called the Android XR Developer Catalyst Program, which will give away 1,000 dev kits to developers and hold a hackathon in the days before AWE 2026 in June.

Google and XREAL have promised a commercial launch for Project Aura this year. At this point, I think Project Aura is ready to sell to consumers and enterprises and may be the first full-AR device to gain traction in the industry (if Snap doesn’t beat them to it). Project Aura is Google’s hedge against Apple’s, Meta’s, and Snap’s AR efforts, and it will probably be a lot more scalable and possibly more affordable. Mind you, I don’t expect this product to be cheap — it could easily retail for $1,500 — but I do think it will be much more competitively priced than the Galaxy XR or the Vision Pro, which is important for both consumers and enterprises.

Google Beam Becomes More Than Just Spatial Video Conferencing

Google Beam is evolving beyond just a spatial video conferencing solution, becoming a high-end 2-D conferencing product as well. For the longest time, Google Beam, now called Dimension by HP, was solely for 1:1 spatial video conferencing with real-time, photorealistic video so real you felt like you could touch the other person. Yet considering the cost of each unit ($25,000) and the monthly subscription fee ($500), it makes a lot of sense that Google wants to make the product useful for more than just spatial conferencing.

At Google I/O, I had the opportunity to test two of Google Beam’s newer experiences, which were surprisingly both in 2-D. The first took advantage of the high resolution and large screen size to enable better virtual presentations and video conferencing. This includes showing multiple people in life-size format with spatial audio to enhance one of the most typical conferencing situations, a multi-user conference call. Google also demonstrated an experimental agentic conferencing experience that allows the user to interact with an avatar powered by Gemini, which speaks dozens of languages and can do many of the things the Gemini app can help with today.

Google and HP are preparing to make the product generally available this year, and pilots with large enterprise customers are already underway. I think that Google is clearly working on making the $25,000-per-seat and $500-per-month license justified with these new conferencing capabilities. I hope that Google continues to add new features and enhance the experience with AI. But that said, I do think the market is still fairly limited to high-net-worth people and C-suite executives who want to cut down on travel. Google Beam is the future — it’s just not that accessible yet.

Google Continues to Blend AI and XR

Google continues to integrate its latest AI capabilities directly into its spatial computing products. Whether it’s the ever-growing assortment of Android XR devices or the Google Beam platform, Google continues to use AI to enhance the user experience. I’ve long said that spatial computing devices are also among the best interfaces for interact with this new AI and agentic era, using natural language and real-world inputs to provide AI with the context it needs to be insightful and personalized for the user. Google has some really promising launches in the spatial computing space coming this year that should pay off some of its investments in AI. Google has also developed Android XR and Gemini in tandem in ways that will be difficult to compete with for Apple, Meta, and Snap. We are finally getting the launches we’ve been anticipating, and now it’s time to see how they perform.