Keynote kicks off a Gemini-first Google I/O 2026
Under clear skies at Shoreline Amphitheatre, Google I/O 2026 opened with a keynote that made one message unmistakable: Gemini is now the connective tissue across Google’s products, platforms, and research. The main keynote, streamed globally at 10 a.m. PT, set the tone with a rapid-fire sequence of Google AI announcements and demos spanning phones, laptops, and wearable devices. On site, Google executives emphasized an “agent-first” future, where Gemini doesn’t just answer questions but takes actions on users’ behalf across apps and services. The conference runs May 19–20, with a follow-up Developer Keynote and a slate of technical sessions and codelabs going live on May 21. From the outset, Google framed the event as both a response to intense competition in the AI space and a statement of intent: Gemini will be everywhere, from search and Workspace to Chrome, Android XR, and beyond.
Gemini 3.5 Flash and Gemini Omni model headline AI upgrades
At the heart of Google I/O 2026 were new flagship models: Gemini 3.5 Flash and the Gemini Omni model. Gemini 3.5 Flash, which quietly appeared for some users earlier as “Gemini 3 Fast,” finally received a formal on-stage debut, positioned as a lean, high-speed model for everyday queries and agentic workflows. Alongside it, Google spotlighted Gemini Omni as the next-generation multimodal system, designed to unify text, image, and video generation within a single architecture. Already rolling out across Google Flow and the Gemini app, Omni is central to Google’s push toward richer, more graphical answers in experiences like the upgraded Google Search bar. While Gemini 3.5 Pro stayed in the background with only hints of what is “coming soon,” Omni and Flash clearly framed Google’s strategy: a spectrum of Gemini capabilities optimized for latency, creativity, and deeply integrated assistant-style behavior.
A redesigned Gemini interface and AI woven into everyday apps
Google’s software story at Google I/O 2026 centered on how Gemini shows up in the products people already use. The Gemini app received what Google called a complete redesign, built in a new visual language known as Neural Expressive. The refreshed UI abandons the traditional wall of text, instead streaming responses in segments with supporting graphics and imagery, similar to Google’s Generative UI experiments. Gemini Live allows users to fluidly switch between typing and speaking, pause the microphone mid-thought, and rely on improved language and dialect support. Across Chrome and Workspace, Gemini becomes more visible and more agentic, surfacing as upgraded search experiences and in-app copilots. Google Flow, now backed by Omni Flash, evolves into a fuller creative studio for video, while the new Pics image tool taps Nano Banana to generate or edit visuals. The throughline is clear: generative, multimodal responses are the default, not the exception.
Spark, AI Studio, and new web tools empower developers
Developers were given a prominent role in the keynote, with Google rolling out fresh tools to build, test, and ship AI-powered experiences. The Gemini desktop app introduced Spark, an agent mode that can work with local folders, connectors, and skills, hinting at a future where developers orchestrate complex workflows through AI rather than manual scripting. Google AI Studio is set to gain a dedicated mobile app, letting creators prototype prompts and even write code from their phones. On the web side, Modern Web Guidance, currently in preview, offers patterns and resources to steer coding agents as they build sites, while new Chrome DevTools for agents aim to streamline debugging and diagnostics. WebMCP, now available, lets developers turn existing pages into agent toolkits with more autonomy. These moves collectively signal a push to make agentic development mainstream, not just an experimental niche.
Android XR, smart glasses, and AI for science expand Gemini’s reach
Beyond phones and laptops, Google I/O 2026 showcased how Gemini will inhabit new form factors and domains. Samsung’s Android XR glasses, built in partnership with Warby Parker and Gentle Monster, offered a first consumer-ready glimpse of intelligent eyewear, with multiple styles targeting different tastes and budgets. Google also confirmed its first audio-based smart glasses arriving in the fall, delivering private Gemini assistance through spoken responses rather than screens. On stage, Google demonstrated AI glasses recognizing a scene, playing music, and guiding a coffee order via DoorDash while the user’s phone stayed in their pocket. In parallel, Google unveiled Gemini for Science, a research-focused initiative using AI tools to track scientific literature, summarize new findings, and build digital twins of Earth for climate and weather simulations. From wearables to scientific discovery, Gemini is positioned as both a personal assistant and an engine for large-scale insight.