Google Cloud (GOOGL) announced a new AI platform dedicated to enterprises, called Gemini Enterprise Agent Platform, dedicated to the selection and construction of artificial intelligence models. The platform replaces Vertex AI and allows teams to build AI agents aimed at companies’ products, services and operations.
With the Gemini Enterprise Agent Platform, technical teams can develop agents and deploy them within the enterprise through the Gemini Enterprise application. This allows technology sectors to guarantee control, governance and security for the tools developed.
In addition to their own development, companies also have access to more than 200 models already developed in Gemini’s own language, as well as support for third-party models, such as those from Anthropic.
Continues after advertising
Almost 75% of Google Cloud customers use AI products in their companies. Over the past 12 months, 330 of them have processed more than a trillion tokens each, while 35 have reached the milestone of 10 trillion tokens — units of data processed — with Google models.
“Gemini Enterprise is now the end-to-end system for the Agentic Era — the connective tissue between data, people, and all applications and agents that transforms processes into a single intelligent flow,” said Thomas Kurian, CEO of Google Cloud. “It’s not about offering individual services in isolation; the company offers a vertically optimized stack, where everything is developed together to deliver the scale and efficiency required by this new era of AI in production.”
Announced during Google Cloud Next ’26, the Gemini Enterprise Agent Platform was accompanied by a series of other new developments from the developer involving AI technology dedicated to corporations.
Big tech also presented the 8th generation of TPUs — processing units — customized for AI training and inference: the TPU 8t and the TPU 8i. Together, the two new chip models deliver performance per dollar greater than 80% compared to the previous generation, according to the company.
