Thirteen announcements. Google's I/O 2026 keynote on May 19 was the company's most AI-dense event to date - covering a new model family, a Search overhaul, subscription updates, hardware, and another attempt at smart glasses. The full list of the day's biggest announcements runs long.
Gemini 3.5: From Answering to Acting
The headliner was Gemini 3.5, a new family of AI models Google described as "frontier intelligence with action." The "action" framing is the important part. It signals Google is building Gemini not just to answer questions but to execute tasks - browsing, completing forms, making purchases, taking steps on your behalf. That positions Gemini 3.5 in direct competition with the agentic direction OpenAI has been building into ChatGPT and Anthropic has been pursuing with Claudee Code](/tools/claude-code/).
Multiple model variants in the family indicate different options optimized for speed versus raw capability - a pattern that has become standard across the industry since OpenAI introduced it with GPT-4 Turbo in 2023. Google also updated its AI subscription tiers at the event, expanding what's available across Google One and the Google AI plans.
AI Mode Is Rewriting Search
AI Mode - which generates a longer, synthesized answer at the top of search results instead of a list of links - moved closer to broad availability at I/O. Google's own framing is that it changes how people search, shifting from short keyword queries toward conversational questions that get direct answers.
For anyone running a website, content business, or online store, AI Mode is the most significant Search development in years. When a user gets a complete answer from a generated summary without clicking through, organic traffic to the underlying source drops. Google benefits from users staying in its interface longer; publishers and website owners depend on that click-through traffic. That tension remains unresolved, and expanding AI Mode pushes further in one direction.
Gmail also received updates at the event, building on AI drafting and summarization features added over the past year.
Project Aura: Smart Glasses, Take Two
Google Glass was announced in 2012, launched publicly in 2013, and was functionally dead as a consumer product by 2015. It became shorthand for tech overreach. Project Aura is Google's next attempt.
The conditions are different now. Meta's Ray-Ban smart glasses have become a genuine product with real users. Wireless earbuds have normalized wearing tech on your face. And AI assistants are capable enough today to justify a dedicated wearable - in 2013, Siri struggled with basic requests and Google Now was six months old. Google showed Aura running Gemini 3.5 at I/O, framing it as a wearable AI companion. No pricing or ship date was announced.