Google I/O 2026 opens tomorrow with a keynote at 10 AM PT on May 19. If you run a business that uses Google Workspace, deploys AI tools, or is trying to figure out what the next wave of enterprise AI looks like, this is worth paying attention to.

Here is what we know heading into the event, what is expected, and why it matters beyond the usual developer fanfare.

What Google Already Confirmed Last Week

On May 12, Google ran The Android Show: I/O Edition — a pre-conference event that gave us confirmed details before the main keynote. The most consequential announcement was Gemini Intelligence.

Rather than positioning Gemini as a separate assistant you open in a tab, Google is embedding it as the intelligence layer running underneath Android itself. That means AI capabilities — summarisation, reasoning, task completion — built into the operating system, not bolted on top of it. The rollout covers phones, tablets, laptops, wearables, and cars.

Google also announced Googlebook: a new line of AI laptops built from the ground up around Gemini Intelligence, made in partnership with Acer, ASUS, Dell, HP, and Lenovo. These are not budget machines. They are positioned as premium business hardware with AI inference built into the device rather than relying entirely on cloud calls.

For teams using Chrome, Gemini integration begins rolling out in late June for Android devices running Android 12 or higher. It will summarise pages, help with editing, and answer questions in context — without switching apps.

What to Expect at the Main Keynote

The headline expectation is a major Gemini model update, widely reported as Gemini 4.0. Google has not officially confirmed the version number, but the direction is clear: a unified multimodal model that handles text, images, audio, video, and code in a single prompt, with a significantly larger context window than the current generation.

From an enterprise perspective, the agentic capabilities are the piece worth watching most closely. Gemini Enterprise Agent Platform — already in limited availability — lets organisations build and manage autonomous AI agents for multi-step workflows. An upgrade here would mean AI that can handle complex business processes end to end, not just answer questions or draft content.

Google also confirmed it will preview Android XR glasses at I/O, with Gemini built in. The near-term enterprise applications are still speculative, but the direction points toward hands-free AI assistance in field operations, warehousing, and service environments.

For developers, an Aluminium OS announcement is expected — a new operating system designed around AI-first interaction patterns.

The Competitive Context

Google is not operating in a vacuum. Anthropic’s Mythos model — currently in a controlled enterprise preview — and OpenAI’s GPT-5.5 are both raising the bar on what enterprise customers expect from frontier AI. Google’s paid Gemini Enterprise users grew 40% quarter-over-quarter in Q1 2026, which is strong momentum, but the pressure to demonstrate that Gemini can compete at the frontier level is real.

The Googlebook hardware play is an interesting counter-move. If Google can make AI inference happen on-device rather than purely in the cloud, that addresses latency, cost, and data residency concerns that enterprise buyers consistently flag as blockers to adoption.

What This Means for Business

A few things to take away heading into the keynote tomorrow:

The AI layer is moving into the OS. Gemini Intelligence embedded in Android means AI tools become default, not optional extras. For any business using Android devices — whether in a call centre, in the field, or in the office — this changes the baseline of what your team can do without new software purchases.

Enterprise agentic AI is becoming a real product category. If Google announces a significant upgrade to its Gemini Enterprise Agent Platform, it joins Salesforce Agentforce, ServiceNow’s autonomous agents, and Microsoft Copilot in a competitive market for AI that does work autonomously. Business leaders evaluating AI platforms will have more to compare.

On-device AI changes the economics. Googlebook laptops and Gemini built into Android hardware reduce dependence on cloud calls. That is relevant to businesses in regulated industries where sending data to third-party cloud APIs creates compliance friction.

Voice and multimodal AI is getting more capable. A unified Gemini 4 model that handles audio natively — combined with the realtime voice models OpenAI launched earlier this month — signals that voice AI employees are moving from novelty to standard operating infrastructure faster than most organisations have planned for.

Tomorrow’s keynote starts at 10 AM PT. We will be watching the Gemini Enterprise announcements, the agentic AI capabilities, and anything that changes how businesses should think about their AI stack in the second half of 2026.