Google has spent the past two years telling the world it is an AI company. I/O 2026 is where it has to prove that the products are catching up to the positioning.
Google I/O 2026 begins May 19, and if the trajectory of the past two years holds, artificial intelligence will dominate nearly every announcement on the agenda. The two-day event will be livestreamed and is expected to cover updates across Gemini, Android, Chrome, and hardware — with keynotes, developer sessions, and what Google is calling “Dialogues” conversations about the direction of AI.
A word of preparation: last year’s keynote ran close to two hours and touched nearly every corner of the company’s product lineup, from consumer apps and developer tools to hardware. This year’s event is unlikely to be shorter.
A New Gemini Model
A major Gemini model update is widely expected to be the centerpiece announcement. Whether it arrives as Gemini 4.0 or an intermediate release, a newer and more capable version of Google’s flagship AI model is almost certainly coming, according to CNET’s preview of the event. Gemini is already embedded across the majority of Google’s product lineup, which means whatever is announced will have immediate reach across Search, Workspace, Android, and beyond.
On the more speculative end, updates to other AI tools — including Nano Banana, Gemma, Lyria, and Genie — are possible, along with a potential Veo 4 announcement for Google’s AI video generation platform. None of that is confirmed. If Veo does receive a significant update, YouTube integration would be the natural next step given how aggressively Google has been pushing AI across its platforms.
New Gemini Live Voice Models
Pre-show leaks suggest a meaningful upgrade is coming to Gemini’s voice capabilities. A hidden model selector in the Google App reportedly reveals seven previously undisclosed Gemini Live voice models currently under internal testing, including variants with codenames such as “Capybara” and “Nitrogen.” One of those models identified itself as “Gemini 3.1 Pro” when queried — a step up from the Flash Live model currently powering Gemini Live. Early testing found meaningful behavioral differences between models, including variations in memory, location access, and fact-checking capability. The infrastructure to switch between models appears to already be built; it is simply not yet public.
New Video Editing Tools
On the video side, a separate leak points to something called Gemini Omni — a new video generation model that has been surfacing for select users ahead of the event. Described as capable of remixing videos, in-chat editing, and template-based creation, Omni appears to be an evolution of Google’s existing Veo foundation. Early demo footage has reportedly shown strong results, though the model appears computationally intensive: one user reportedly consumed 86 percent of their daily AI Pro plan allowance generating just two short clips.
Also Read: Inside Google’s $175 Billion Bet on the Agentic Enterprise
Aluminium OS
Google introduced the Googlebook — its new AI laptop category — at The Android Show event this week, with company representatives confirming the devices will not run ChromeOS but rather an as-yet-unnamed operating system. The strong implication is that Google’s long-in-development project to create an Android-based operating system for laptops — internally referred to as Aluminium OS — is expected to receive significant airtime at I/O.
The project would merge Android and ChromeOS capabilities into a single platform targeting a broader consumer laptop audience. Android Ecosystem President Sameer Samat confirmed earlier this year that the platform remains on track for a 2026 debut, pushing back against court documents from the Google antitrust trial that suggested a possible 2028 timeline. ChromeOS is not going anywhere — Samat characterized the two as parallel strategies rather than one replacing the other.
Android XR Glasses
After debuting as a concept at last year’s I/O, Google confirmed in December 2025 that its Android XR smart glasses would launch in 2026. Two distinct products are in development under that umbrella.
The first is a display-free pair of AI glasses equipped with a camera, speakers, and microphones for hands-free Gemini interaction — a form factor that draws direct comparisons to Meta’s Ray-Ban glasses. The second, more ambitious product adds an in-lens display capable of surfacing navigation directions, translation captions, and other information visible only to the wearer. Both are being developed in partnership with Samsung, Gentle Monster, and Warby Parker, and both run on Android XR — the same operating system powering the Samsung Galaxy XR headset that launched in October 2025.
A working demo of both products at I/O 2026 would be a meaningful step forward. What remains unclear is when the display version will actually ship. Google has confirmed the simpler AI glasses are expected this year but has not given a launch date for the display model.
Android 17
Google has been rolling out Android 17 developer betas since February, with a final release expected around June or July ahead of the Made by Google Pixel hardware event in August. The beta has been relatively light on headline features so far, with app bubbles — a floating window system for quick application access — among the more notable additions confirmed to date.
Also Read: Why Most AI Pilots Never Make It to Production
More Agentic AI
Across the board, expect a significant push for agentic AI — systems capable of handling multi-step tasks on a user’s behalf with minimal ongoing input. Google has been signaling this direction for some time, and I/O is the natural stage to demonstrate what that looks like in practice through Gemini. OpenAI is separately rumored to be building a phone concept around agentic AI, suggesting the competitive pressure on Google to show concrete progress is considerable.
The Future of Search
Google Search updates are expected as well, with announcements likely around AI Mode and AI Overviews — the features that have most visibly changed what it means to use Google as a search engine. The tension between Google’s identity as a search company and its ambitions as an AI platform will be visible in every announcement made on May 19.
