Google yesterday unveiled new products at its Google Cloud Next 2025 conference, including enhancements to its Google Cloud artificial intelligence (AI) Hypercomputer stack, new supercomputing and AI capabilities added to its cloud services, and an upgrade of its Gemini generative AI chatbot.
Addressing the audience at the event, held in Las Vegas, Sundar Pichai, CEO of Google and Alphabet, said the tech giant has been investing in AI and machine learning for more than two decades. This is because it is “the most important element” to help the company advance its mission: to organise the world’s information and make it universally accessible and useful.
In 2024, Google delivered 3 000 product advancements across Google Cloud and Workspace. There are now over four million developers building applications with the help of Gemini, the company’s advanced AI model family, he noted.
Pichai noted this is coupled with a 20-times increase in Vertex AI usage in the past year, driven by the adoption of Gemini, Imagen (image generation model) and Veo (video generation model).
“A couple of weeks ago, we released a new model, Gemini 2.5, which is a thinking model that can reason through its thoughts before responding. It’s our most intelligent AI model.
“Gemini 2.5 Pro is now available for everyone in Google AI Studio, Vertex AI and in the Gemini app. We’re also announcing Gemini 2.5 Flash, our low latency and most cost-efficient thinking model. With 2.5 Flash, users can control how much the model reasons, and balance performance with their budget,” explained Pichai.
Today, 15 of the company’s product offerings − including seven with two billion users − are using Gemini models. AI deployed at this scale requires inference, which enterprises can benefit from, to build their own AI-powered applications, Pichai stated.
Workspace provides over two billion AI assistants monthly to business users, across the globe, he added.
“All of this is powered by our global infrastructure, which has grown to 42 regions, with new locations in Sweden, South Africa and Mexico, and rapid expansion under way in Kuwait, Malaysia and Thailand.
“These regions are connected by more than two million miles of terrestrial and subsea cables, and have more than 200 points of presence across 200+ countries and territories, creating a resilient foundation for the AI-powered future.”
Google’s AI Hypercomputer is a supercomputing system designed to simplify AI deployment, improve performance and optimise costs, says the company.
Pichai introduced updates in the AI Hypercomputer stack, including AI-optimised hardware, software updates for inference and flexible consumption options to help businesses control costs.
Also speaking at the event, Thomas Kurian, CEO of Google Cloud, pointed out that itswide area network makes its global private network available to all Google Cloud customers.
He said Google Cloud WAN is a fully-managed enterprise backbone for enterprise WAN architectures, with a claimed improvement of up to 40% in network performance.
“New advancements in Google Distributed Cloud lets customers bring Gemini models on-premises and allow support for a full suite of generative media models and Gemini 2.5, our thinking models. Innovations in Vertex, like Agent Development Kit and Agent2Agent Protocol, enable a multi-agent ecosystem,” added Kurian.
“[Google has also introduced] enhancements to Agentspace [storage solution] that let employees benefit from AI, and a number of announcements across Workspace, databases, analytics, cyber security, our ecosystem and more.”
Share