Google Cloud Makes Key Agentic AI Announcements at Next ‘26

To coincide with the Google Cloud Next ‘26 conference – Google Cloud’s largest event – the cloud giant is making a number of announcements, including the new Gemini Enterprise Agent Platform, advancements to its AI Hypercomputer architecture, and introducing the Agentic Data Cloud.

A single platform for agent development, orchestration, and governance

Google Cloud is introducing the new Gemini Enterprise, an end-to-end system for the Agentic Era, billed as the “connective tissue between your data, your people, and all of your apps and agents that transforms all of your processes into a single, intelligent flow.”

The organization is expanding the Gemini Enterprise portfolio with the Gemini Enterprise Agent Platform, the Gemini Enterprise app, and an open partner ecosystem.

Vertex AI evolves into a full-scale developer platform

The Gemini Enterprise Agent Platform is the organization’s new developer platform and an evolution of its Vertex AI. 

The platform brings together Google Cloud’s full suite of models, development, and tuning services and offers new features to help businesses build, scale, govern, and optimize agents that can execute complex business workflows autonomously.

The new features allow organizations to:

• Build: Organizations can select the right environment for the job, from the low-code, visual interface of the new Agent Studio, to the code-first logic of the upgraded Agent Development Kit (ADK). Google Cloud has simplified the lifecycle with AI-native coding capabilities to help ship production-grade agents faster.
• Scale: Access to a clear path to production with the re-engineered Agent Runtime. It supports long-running agents that maintain state for days at a time and are backed by Memory Bank for persistent, long-term context.
• Govern: Establish centralized control through Agent Identity, Agent Registry, and Agent Gateway to ensure every agent – whether built on the Agent Platform or sourced from our partner ecosystem – has a trackable identity and operates within enterprise-grade guardrails.
• Optimize: Ensure quality with Agent Simulation, Agent Evaluation, and Agent Observability to provide full execution traces and a real-time lens into agent reasoning, helping ensure agents always hit goals.

“What we’re evolving the platform to do … is to connect to systems of record … Agents are  then built, and we give them identity, security, telemetry, and governance to manage them,” said Thomas Kurian, CEO of Google Cloud. “We provide a scalable agent execution to orchestration environment, and then we provide tools to evaluate and optimize the quality of these agents.”

Gemini Enterprise app integrates governance, productivity, and third-party systems

The Gemini Enterprise app will allow teams to discover, create, share, and run AI agents in a single, secure environment. 

It is built on the Agent Platform, which offers all governance, security, and identity capabilities while integrating with an organization’s enterprise data. This includes data in third-party systems through connectors.

Further, the open partner ecosystem enables the discovery and deployment of a wide range of third-party agents from organizations such as Oracle, Salesforce, and ServiceNow within a secure, governed environment.

Google Cloud’s AI Hypercomputer architecture

The organization is combining fragmented components into a unified, agent-native engine.

By advancing its AI Hypercomputer architecture with new hardware and software, it can handle the massive scale and low-latency requirements of the agentic era. 

Among these advancements are:

• The eighth generation of Google’s custom Tensor Processor Unit (TPU) with two purpose-built architectures for training and inference: TPU 8t and TPU 8i.
• Google Cloud Virgo Network: A new megascale AI datacenter fabric to meet the scale and latency requirements of the AI era.
• New storage innovations, including Managed Lustre which now delivers 10 TB/s of bandwidth, a 10x improvement over last year and up to 20x faster than other hyperscalers.

Updates to the Agentic Data Cloud

Google Cloud is also introducing Agentic Data Cloud, an AI-native architecture that transforms the enterprise data platform from a static repository into a reasoning engine.

The Agentic Data Cloud is a “System of Action” that is built for agent scale.

Among the features of the Agentic Data Cloud are:

• A universal context engine that provides agents with trusted business context to drive higher accuracy.
• Agentic-first practitioner experiences to evolve the role of data practitioners and developers as orchestrators of agents.
• An AI-native, cross-cloud lakehouse to eliminate data silos by connecting your entire data estate.

On top of this, Google Cloud is announcing four major performance breakthroughs:

1. Lightning Engine for Apache Spark to deliver twice the price-performance over the proprietary market alternative.
2. Managed Lustre for delivering up to 10 terabytes per second of throughput to make sure data moves quickly enough for demanding models.
3. Bigtable will now support an in-memory tier that delivers submillisecond read latency for real-time applications, eliminating the need for separate, side-by-side caching layers.
4. BigQuery fluid scaling will help lower costs by up to 34 percent on average for autoscaling workloads, scaling resources up instantly when agents act and back down when they don’t.

“All of this innovation is propelling customer adoption. In fact, nearly 75 percent of Google Cloud customers are using our AI products to power their businesses,” said Kurian. 

“In addition, over the past 12 months, 330 Google Cloud customers each processed more than one trillion tokens, while 35 reached the 10-trillion-token milestone with our models. Our first-party models now process more than 16 billion tokens per minute via direct API use by our customers, up from 10 billion last quarter.”