Google I/O 2026 kicks off Tuesday, May 19, with the main keynote starting at 10 a.m. Pacific Time at the company’s annual developer conference in Mountain View. The two-day event runs through May 20 and is free to watch online via Google’s YouTube channel and the I/O website, so nobody needs a badge to follow along.
Google has framed this year’s conference around what it calls the “agentic era” of development, a phrase from the company’s developer blog that signals the keynote will lean heavily on AI systems capable of handling multi-step tasks with minimal human hand-holding. That framing puts Gemini at the center of the show, and previews from several technology outlets suggest the AI model family will be the through line connecting nearly every product announcement on the agenda.
Android 17 is the other headliner everyone is watching. The next major version of Google’s mobile operating system is widely expected to get stage time, though the specifics of what changes it brings remained unconfirmed before the keynote. Google has not released a public timeline for Android 17, and I/O has historically been where the company lays out its roadmap for developers who need to prepare their apps for whatever comes next.
Then there are the glasses. Google has confirmed plans to release smart glasses built around Android XR in 2026, according to Mashable, which noted that the company has not settled on a final product name. Wired said attendees and viewers should expect another look at the Android XR smart glasses, which have been previewed before but not yet shipped. CNET framed the wearable question more broadly, asking whether smart glasses could become a genuine pillar of Google’s device lineup rather than a niche curiosity. That is a reasonable question given how the category has underdelivered for years, and Google’s answer, or at least its pitch, is apparently coming today.
PCMag said before the keynote that it expected significant changes to Gemini and Android Auto, along with possible details on a device it referred to as a Googlebook running something called Aluminum OS. That item was firmly in the realm of pre-show speculation rather than confirmed news, and it is the kind of thing that either gets announced or quietly disappears from the conversation by Tuesday afternoon.
Beyond hardware, Google’s developer blog described the event as a chance to showcase tools designed to automate complex workflows and simplify building AI-ready applications. That language is aimed squarely at the developer community, and there is typically a second keynote focused specifically on APIs, tools, and features that engineers can actually use. Mashable noted that this developer-focused session usually runs separately from the main consumer keynote, giving technical attendees something more substantive than product trailers.
Watching options are broad by design. The keynote livestreams on Google’s official YouTube channel and at the I/O site. CNET planned a YouTube viewing party with live updates from reporters on the ground in Mountain View, plus coverage before and after the main event. Canaltech reported that Google will also offer an American Sign Language version of the livestream alongside the standard broadcast, which is a straightforward accessibility measure worth mentioning for anyone who needs it.
Chrome and Cloud were also listed by Google as areas that would see updates across the two-day schedule, though details ahead of the keynote were thin. Broader Gemini integrations across Google’s product suite were another widely cited expectation, consistent with the company’s push over the past year to embed AI into Search, Maps, Workspace, and essentially everything else it ships.
The conference arrives at a moment when Google is under genuine competitive pressure in AI, with rivals moving fast and the search business facing structural questions it has not had to answer before. I/O is as much a statement of direction as it is a product showcase, and the choice to lead with an “agentic era” keynote suggests Google wants to make clear it sees the next phase of AI as less about chatbots answering questions and more about systems that take action. Whether the announcements live up to that framing is what the next several hours will determine.