The pace of technological change since last year’s Cloud Next has never been faster, and Google Cloud has incredible momentum.
Our first-party models now process more than 16 billion tokens per minute via direct API usage by our customers, up from 10 billion last quarter. To support and drive this growth, just over half of our total machine learning investment in 2026 is expected to go to the Cloud business to benefit our cloud customers and partners.
You can read all about our momentum and the extraordinary range of partnerships and innovations we’re announcing at Cloud Next.
I will just highlight four key areas.
1. We are firmly in the age of agentic Gemini
Last fall, we introduced Gemini Enterprise, the end-to-end system for the agency era—the connective tissue between your data, your people, and your goals.
It has great momentum: in Q1 we saw 40% growth in paid monthly active users quarter over quarter.
Through this rapid growth, we have seen how any employee in any organization can become a builder. This is an incredible shift, but it comes with complexity. The conversation has gone from “Can we build an agent?” to “How do we manage thousands of them?”
That’s why we’re introducing our new Gemini Enterprise Agent Platform. It provides the secure, full-stack connective tissue you need to build, scale, manage, and optimize your agents with confidence—mission control for the agent business.
2. Using artificial intelligence to defend against security threats
Although AI can increase security risks, our Cloud customers now have AI on their side to protect their organizations. Today, we’re unveiling a range of new threat detection agent solutions as part of an AI-powered cybersecurity platform that combines Google’s Threat Intelligence and Security Operations with Wiz’s Cloud and AI Security Platform.
In addition, we are launching Wiz’s new AI Application Protection Platform (AI-APP), which provides autonomous protection, from code to cloud to runtime, across multicloud, hybrid and AI environments.
3. Introducing our eighth generation TPUs
In the era of AI agents, the infrastructure must evolve to handle the most demanding AI workloads. This year we bring the eighth generation of our Tensor Processing Units with a dual chip approach:
- TPU 8hoptimized for training, scales up to 9,600 TPUs and 2 petabytes of high-bandwidth shared memory in a single superpod. It achieves three times the processing power of Ironwood and delivers up to 2x more performance/watt.
- TPU 8ioptimized for inference, connects 1,152 TPUs in a single pod, dramatically reducing latency, with 3x more on-chip SRAM, to deliver the massive throughput and low latency needed to cost-effectively run millions of agents simultaneously.
We will offer these to Cloud customers as a core part of our range of compute processors alongside a portfolio of NVIDIA GPU instances. Read more in our blog post.
4. Stay ahead as “Kundenul”
To be the best partner, we always want to be “customer zero” for our own technologies. This helps us imagine, test, build and scale the best Google technologies for our cloud customers for today and tomorrow. Our database service Bigtable, which powers so many Google services, and our TPUs, which have been so important in training and running our Gemini models, are good examples.
Here are a few more recent ones:
First, coding.
- We’ve been using AI to generate code internally at Google for a while. Today, 75% of all new code at Google is now AI-generated and approved by engineers, up from 50% last fall.
- We are now switching to truly agentic workflows. Our engineers orchestrate fully autonomous digital task forces, fire agents, and accomplish incredible things.
- Recently, a particularly complex code migration performed by agents and engineers working together was completed six times faster than was possible a year ago with engineers alone.
- And with our recent launch of the Gemini app on MacOS, the team built the first release with our agentic development platform Antigravity, going from an idea to a native Swift app prototype in days.
Other, security.
- We have long led the industry in security. Now our Security Operations Center agents automatically triage tens of thousands of unstructured threat reports each month, reducing threat mitigation time by more than 90%. And we’ve built and actively use Gemini-based AI agents (like CodeMender) to find and, importantly, fix critical software bugs.
Third, ours operations.
- For the launch of Gemini in Chrome, our marketing teams used our models to quickly generate thousands of variations of our creative assets, which historically would take weeks. Using artificial intelligence led to 70% faster revenue and a 20% increase in conversions, allowing us to market faster and more efficiently.
Congratulations to our Google Cloud team and a big thank you to our partners who are building the future with us. We’ll have a lot more to share about how we’re bringing the latest technology to everyone at Google I/O on May 19.