We're back again for Google Cloud Next – this year we'll have two days of live coverage coming straight from Google Cloud's biggest annual event.
The first day kicked off with the main keynote covering tens of separate announcements. The rest of the conference provided opportunities to explore them in more detail so we can bring you all the latest news from Google Cloud this year.
The venue is filling up with around 32,000 attendees as Google Cloud gets ready hold its biggest event of the year, packed with new product announcements and more.
Good morning from Google Cloud Next. The first keynote kicks off at 9am PT, so join us in a little over an hour for all the details as they roll out.
In true tech conference style, a live DJ set is warming the audience up ahead of today's announcements
"How do you move AI into production cross your entire enterprise?" Google Cloud CEO Thomas Kurian asks as he welcomes the audience.
Spoiler alert, he's already given us the answer – it's a unified stack. That's what we'll be hearing about over the next two days.
Around 75% of all new Google code is AI-generated, Google CEO Sundar Pichai says in a pre-recorded video. Around a year ago, we reported that AI was accounting for around 25% of Google's new code (and a third for Microsoft).
The focus of today's keynote and this year's announcements will be Google Cloud's new Gemini Enterprise Agent Platform, described as an "end-to-end system for the agent era"
We're also hearing about Gemini 3.1 Pro, Gemini 3.1 Flash Image (Nano Banana 2), Veo 3.1 Lite and Lyria 3 Pro
And name drop... Google Cloud is reminding us about the landmark Apple partnership, which will drive this year's much-improved Siri
Name drops are continuing as Thomas Kurian reminds us that Google Cloud partnered with NASA/Artemis II on flight readiness with agentic AI
Projects in Gemini Enterprise, coming soon, serves as a shared workspace for teams and agents
"The era of the pilot is over, the era of the agent is here," Kurian declares as he welcomes colleague Erica Chuong for a demo.
For another live demo, 3x Olympic Gold Medallist and snowboarder Shaun White is showing us how Google Cloud and Gemini can analyze the details of a video for detailed analytics.
Learning a trick on the mountain is one thing, White says, but visualizing and analyzing the physics unlocks something that's never been possible before.
Oh, and of course, the convention center filled with "snow" for White's entry.
Next up, TPU 8t and TPU 8i are being announced
TPU 8t promises 3x performance per pod, support for up to 9,600 TPUs via 3D torus topology, 2 PB of shared bandwidth memory (enough to hold the entire digital collection f the Library of Congress 100x over)
As for TPU 8i, there's now support for up to 1,152 TPUs per pod via Boardfly Topology. We've broken down the "memory wall", Google says
Of course, we've got the fuller details for Google's two new TPUs coming later on
Chief Product and Business Officer Karthik Narain is here for the next launches, kicking off with the Knowledge Catalog, Smart Storage for agent-ready data with instant tagging and zero manual data engineering, and Deep Research Agent
Sticking on the data side, Google is launching the Cross-Cloud Lakehouse that's "completely borderless"
"No more vendor lock-in, just freedom," Narain declares
Just weeks after we learned that Google Cloud completed its acquisition of Wiz, we're now getting an insight into what this acquisition means for Google Cloud.
In a quick demo on stage, we see Gemini Enterprise for Customer Experience handle complex workflows with product queries and even language changes. For the consumer it means that conversational AI can be used via phone support, with the AI agent having access to all the necessary resources behind the scenes.
Yulie Kwon is on stage now to announce the general availability of Workspace Intelligence – deeper context for information gathering with situational awareness and personalization. It turns fragmented information into a clear path forward, she adds.
Additionally, Rapid Enterprise Migration, in Preview, is now up to 5x faster to move an entire organization from M365 thanks to big work on interoperability
Next up, we're heading to a breakout to hear what's new with Gemini, from Google DeepMind
While that's it for the live keynotes, we're still on the ground jumping between breakouts and speaking to Google execs for all the latest details on this year's long list of announcements.
Good morning from the second of two days at Google Cloud Next 2026. This morning, we're having an exclusive session with Google Cloud CEO Thomas Kurian to hear about some of the announcements in greater detail.
In a room of journalists, Kurian is speaking about Google Cloud's European sovereignty principles, model distillation, the increasing pace of product announcements and internal development, and quality control.
Google Cloud is very proud of its $32 billion acquisition of Wiz. Kurian is telling us about how its technology will help to automate cybersecurity – AI can simulate red teams (attack simulation), blue teams (key to identifying the vulnerabilites), and green teams (for figuring out the fixes).
Speaking with TechRadar Pro in an exclusive interview, Google Cloud Global Director of Global Healthcare Solutions Aashima Gupta reiterated a message we've been hearing a lot this year, not just at Google Cloud Next, but also from Nvidia CEO Jensen Huang (speaking at Adobe Summit earlier this week).
Artificial intelligence will handle the chore, she explained, while workers will be able to focus on the actual purpose of their jobs.
For healthcare clinicians, this means they'll be able to spend more time taking care of patients – but we'll explore this in greater detail separately.
Speaking with EMEA journalists in a breakout about defining ROI in the world of agentic AI, Oliver Parker, VP for AI GTM and Alex Rutter, EMEA MD for AI explain that it's difficult to standardize this. Instead, each organization should consider what they want from AI as the process becomes far more personalized.
In about half an hour, we'll be getting a hands-on, real-time demo of some of the new AI tools in Google Workspace to see how Google Cloud has been tackling some real use cases
Now to see some of the latest additions to Workspace in practice!
We're seeing Workspace Intelligence at play, pulling context from multiple places across Drive, Gmail and more to generate new documents, using the same formatting as a reference document.
Gemini can also address comments from co-workers in Workspace apps like Docs – actions are taken and produced in a 'preview' in-line for the user, who can go ahead to accept or reject them.
Another demo, this time in Sheets, shows Gemini being able to fill out cells based off context, with no additional prompts needed
Next up, we're seeing how a prompt can be entered into the Drive search bar to generate an AI Overview
A quick hop and a jump now over to a 'chip chat' session with Mark Lohmeyer (VP/GM, AI and Computing Infrastructure) and Leo Leung (Director of Product Management AI & Computing Infrastructure) to hear about Google's new eighth-gen TPUs this year. We'll be covering these separately soon
Next up, it's a hands-on demo for us to build an agent with Gemini Enterprise
That's more or less a wrap on our time at Google Cloud Next, but keep an eye out on what's to come as we break down some of the product announcements, including the eighth-gen TPUs and new Workspace improvements. Thanks for joining us!
