Tech Landscape: Google I/O 2026
The headlines from the keynote of Google’s annual developer event
Hello!
I’m back from my break (Copenhagen, I loved it) with an unusual mid-week edition just to cover the news from this year’s Google I/O. There was a ton of announcements as usual, so I’m just going to highlight a few below. I’ve had to write this very quickly so I apologise for the brevity and any mistakes. The regular newsletter will be back on Monday.
Right, let’s get on with it. Hope you’re well!
Key Technologies
Gemini Omni Flash is the first in the new Gemini Omni family of multimodal media models. Right now it can only generate videos, with improved world understanding (such as physics) and advanced editing capabilities, but future versions will create “any output from any input”. It’s being positioned as ‘the Nano Banana moment for video’, and is available now in the Gemini app and Flow to paid subscribers, and in YouTube Shorts and the YouTube Create app to everyone.
This is the only model I’ve had time to quickly test ⬇️; it’s not a huge leap in terms of visual quality over Veo 3.1, but promises more control and coherent responses. It’s not as powerful as current champion Seedance 2.0, but is more versatile.
Gemini 3.5 Flash is the first language model in the 3.5 series (3.5 Pro will follow later). Google says it excels at agentic workflows, coding, and interactive Web UI and graphics, and is faster than similar rivals - although much more expensive than Gemini 3 Flash. It’s generally available, including in the Gemini app and AI Mode in Search.
Antigravity 2.0 is a fully redesigned desktop app for builders to orchestrate workflows using the Antigravity agentic platform.
blog.google/…/google-io-2026-developer-highlights/
(I wouldn’t usually highlight this but it will come up surprisingly often below.)
The Gemini App
Google says over 900 million people now use the Gemini app, up from 400 million this time last year.
The app has been redesigned with the new Neural Expressive design language. Responses will be less text-heavy, with more illustrating infographics and videos, and Gemini Live will be more tightly integrated in conversations. It’s generally available now, along with the Gemini 3.5 Flash and Gemini Omni Flash models.
Gemini Spark is an agentic mode in the Gemini app which can run complex multi-step tasks in the background. It’s powered by Gemini 3.5 Flash and Antigravity and is integrated with Workspace apps. It powers Daily Brief, “a personalised morning digest”. Spark will be available to Ultra subscribers in the U.S. (and there’s a new mid-tier Ultra subscription available for $100 a month).
blog.google/…/next-evolution-gemini-app/
Search
Google Search will see “the biggest upgrade to our Search box in over 25 years”, redesigned with intelligent suggestions and expanding for more complex search queries. AI Overviews and AI Mode will be more tightly coupled. This will come to all users soon.
Users will be able to set up “information agents” that run in the background to return detailed and ongoing updates, and even code their own apps that will create custom UI to illustrate search results (powered by Gemini Flash 3.5 and Antigravity). These agentic features will come to everyone this summer, and Ultra subscribers in the U.S. will be able to build apps that remain persistent across search sessions.
Shopping is also getting an agentic upgrade, with a Universal Cart that will store products found in Search, Gemini, and YouTube, actively monitoring for price and stock updates. The Agent Payment Protocol (AP2) will create a “digital paper trail” for tracing purchases made with agents (in case your agent buys something you didn’t want). The Universal Commerce Protocol, a standard for agentic shopping, will come to more countries including the UK, and with new partners including Amazon and OpenAI on board.
blog.google/…/google-shopping-cart/
Creative Tools
Google Flow gained the Gemini Omni Flash video model, a new Agent mode that can plan and reason more complex tasks including batch content generation, and Tools for creating bespoke tools and workflows, such as graphical filters and shaders. Google Flow Music added precise section editing and the ability to reinterpret generated songs. Both platforms will get native mobile apps.
Google Pics is a new Workspace app for image creation and editing, with segmentation for easier editing; it will work standalone or in Workspace apps including Slides and Drive. It’s coming soon.
workspace.google.com/products/pics/
Intelligent Eyewear
The first audio glasses with Gemini embedded, designed by Gentle Monster and Warby Parker and built by Samsung and Qualcomm, will launch later this year.
blog.google/…/android-xr-io-2026/
The Android Show
Last week, ahead of I/O, Google announced a series of updates coming to Android. You can watch the full event or a short recap to learn more.
Gemini Intelligence is a new system layer that lays the groundwork for Android to transition “from an operating system into an intelligence system”. Key features include handling multi-step tasks across different apps using screenshots to provide powerful context, completing Web forms with Gemini in Chrome, a new Gboard feature called Rambler which will make sense of unstructured voice inputs, and “vibe coding” custom widgets. blog.google
There’s a useful developer-focused recap to learn more about how it all works.
Creator-focused features for Android 17 include Screen Reactions, which can record yourself and your screen simultaneously, camera and editing features in Meta’s Edits and Instagram, and the tease of Adobe Premiere coming to the platform. blog.google
Googlebook is a new laptop category bringing together Android, Chromebook, and Gemini Intelligence, with Magic Pointer that uses the position of the mouse pointer to provide context to voice instructions. blog.google
Magic Pointer looks very promising as a way to use “this and that” rather than explicitly describing something. You can try a couple of demos for yourself: Maps, Image Edit.
Improved sharing and connection with iOS including Quick Share compatibility with AirDrop coming to more devices, and encrypted RCS messages. blog.google