Google Cloud Next 2026 is underway in Las Vegas this week, and the announcements are coming fast. From brand-new AI chips to a completely rebuilt Gemini platform for businesses, Google used its biggest enterprise event of the year to make its most aggressive push yet into the AI agent era.
If you missed the keynote or just want a clean summary of what was announced, here is everything that matters from Google Cloud Next 2026 — and what it means for you.
What Is Google Cloud Next 2026?
Google Cloud Next is Google’s annual enterprise technology conference. This year it runs April 22 to 24 in Las Vegas, with keynotes from CEO Sundar Pichai and Google Cloud CEO Thomas Kurian.
The central theme of this year’s event is what Google calls “the agentic cloud” — a vision where AI agents do real work inside businesses, not just answer questions. Virtually every announcement ties back to this idea.
TPU v8: Google’s New AI Chips
Google opened its keynote with a major hardware announcement: the eighth generation of its Tensor Processing Units, known as TPU v8. Crucially, this generation introduces two separate chips for the first time — one optimised for training AI models, and one designed specifically for inference (running AI models in production).
Google argues this dual-chip approach delivers better per-token economics than general-purpose AI chips, making it more cost-efficient to run AI agents at scale. This is a direct response to Nvidia’s dominance in AI hardware, with Google positioning purpose-built chips as the smarter long-term choice for enterprises running agent workloads.
Google also announced the Virgo Network, described as a new megascale data centre fabric that underpins its AI Hypercomputer infrastructure and is designed to support the next decade of machine learning workloads.
Gemini Enterprise Agent Platform
The headline software announcement is the Gemini Enterprise Agent Platform — Google’s new end-to-end system for building, scaling, governing, and optimising AI agents inside a business.
The platform includes a new Agent Designer tool for creating agents without code, an Inbox for managing ongoing agent activity, support for long-running agents that operate over hours or days, and a new Skills system that lets agents learn reusable capabilities.
Google framed this as more than a tool suite. The company described it as the operating system for the agentic enterprise — the infrastructure layer that businesses will build on as they shift from AI assistants to AI workers.
Workspace Intelligence
Google also announced Workspace Intelligence, a new layer that runs across all of its productivity apps. Unlike basic AI integrations that pull data from a single source, Workspace Intelligence understands semantic relationships between your Docs, Slides, Gmail, active projects, collaborators, and your organisation’s domain knowledge — all at the same time.
The goal is to give Gemini real contextual awareness when working inside Google Workspace, rather than treating each app as a separate silo. Google is already rolling out ten new Workspace updates connected to this system, including significant enhancements to Gemini in Sheets.
Gemini-Powered Auto Browse for Chrome
One of the most consumer-relevant announcements is Gemini Auto Browse for Chrome Enterprise — a feature that lets Gemini autonomously complete tasks inside the browser, including form filling, web research, data entry, and meeting scheduling. We covered this in detail separately; it is a big deal for anyone who uses Chrome at work.
This is connected to a broader pattern Google has been building for months. Earlier this year, Gemini was rolled out to Chrome users in India, and Gemini Live received a major update adding mid-briefing follow-up questions. Auto Browse is the most powerful Chrome-based Gemini feature yet.
New AI Models: Gemini 3.1, Veo 3.1, and Lyria 3
Google also announced a new wave of model releases at Cloud Next. The lineup includes:
- Gemini 3.1 Pro — the latest update to Google’s flagship reasoning model
- Gemini 3.1 Flash Image (Nano Banana 2) — an updated image generation model
- Veo 3.1 Lite — a lighter version of Google’s video generation model
- Lyria 3 Pro — an upgraded music generation model
These models are rolling out to Google Cloud customers and developers via the Vertex AI platform.
Google’s AI Cybersecurity Platform
Google also used Cloud Next to announce a new AI-powered cybersecurity platform that combines Google Threat Intelligence, Google Security Operations, and the Wiz cloud security platform — following Google Cloud’s recent acquisition of Wiz. The platform includes new agentic tools for automated threat detection and response.
AI Now Writes 75 Percent of Google’s Own Code
Perhaps the most striking data point of the entire event came from Sundar Pichai, who confirmed that approximately 75 percent of all new code written at Google is now generated by AI and reviewed by human engineers. That number was around 25 percent just a year ago. It is one of the clearest signals yet of how rapidly AI is changing how software gets built — even at the companies building the AI.
For context, Google’s AI ambitions extend well beyond the cloud. Gemini Nano 4 is coming to Android devices, and Google is already building a native Gemini app for Mac. If you want to see where all of this fits in the competitive landscape, our Google Gemini vs ChatGPT vs Copilot comparison for 2026 breaks it all down.
Google’s next major consumer event is Google I/O, scheduled for May 19.
Frequently Asked Questions
What was announced at Google Cloud Next 2026?
The key announcements include TPU v8 dual AI chips, the Gemini Enterprise Agent Platform, Workspace Intelligence, Gemini Auto Browse for Chrome, new Gemini 3.1 and Veo 3.1 models, an AI cybersecurity platform, and the Virgo Network data centre fabric.
When is Google Cloud Next 2026?
Google Cloud Next 2026 runs April 22 to 24 in Las Vegas. Keynote replays are available for free on Google Cloud’s YouTube channel and via a free digital pass on the event site.
What is the Gemini Enterprise Agent Platform?
It is Google’s new platform for businesses to build, deploy, and manage AI agents. It includes a no-code Agent Designer, an Inbox for agent activity management, long-running agent support, and a Skills library for reusable agent capabilities.
What are TPU v8 chips?
They are Google’s eighth-generation custom AI processors. For the first time, Google has split the design into two chips — one for training AI models and one for running them in production — arguing this delivers better cost efficiency than general-purpose GPUs.
Is Google Cloud Next different from Google I/O?
Yes. Google Cloud Next is focused on enterprise and developer products. Google I/O is Google’s consumer-facing developer conference and is where iOS, Android, and consumer AI features tend to get their big reveals. Google I/O 2026 is scheduled for May 19.
Conclusion
Google Cloud Next 2026 is one of the most significant enterprise AI events in recent memory. Between the TPU v8 chips, the Gemini Enterprise Agent Platform, Workspace Intelligence, and Auto Browse for Chrome, Google is making a clear case that the agentic AI era has arrived — and that its cloud is where businesses should build it.
The next major moment on the calendar is Google I/O in May, where consumer-facing versions of many of these technologies are expected to be unveiled.
Leave a Reply