Gemini Enterprise Agent Platform
Google's enterprise platform for building and governing AI agents
- Overview
- Formerly Vertex AI, Gemini Enterprise Agent Platform — rebranded at Google Cloud Next 2026 — is Google Cloud's consolidated runtime for building, deploying, and governing production AI agents. It bundles the Agent Development Kit, Agent Studio (low-code), 200+ foundation models via Model Garden (including Gemini and Claude), a managed agent runtime with persistent Memory Bank, AutoML, and MLOps pipelines under one pay-per-use platform. Existing Vertex AI workloads, SDKs, and APIs continue to run unchanged under the new branding.
- Why it matters
- The Vertex AI → Gemini Enterprise Agent Platform rebrand is not cosmetic: Google has inverted the product hierarchy so that agents are now the primary organizing concept, with model training and AutoML demoted to sub-menus. This signals Google's bet that enterprise AI spend will shift from standalone model API calls to orchestrated, multi-step agent workflows — and locks GCP customers deeper into a unified billing surface that spans inference, retrieval, memory, and compute. Teams should watch for pricing step-ups as introductory Gemini 3.x Flash rates are scheduled to double in early 2027, and for growing multi-cloud friction as the platform tightens around Google's own A2A and ADK toolchain.
- In practice
- Teams use the platform to deploy RAG pipelines (Vertex AI Search + RAG Engine) for internal knowledge bases, build customer-facing conversational agents via Agent Studio, and fine-tune Gemini models against proprietary data using supervised fine-tuning jobs. The Memory Bank feature is seeing adoption for financial and CRM use cases where agents need to recall user context across sessions. The main limitation is cost unpredictability: agents stack multiple meters simultaneously (tokens, compute, search queries, grounding), so production costs regularly exceed initial estimates without careful instrumentation.
- Best for
- GCP-native engineering teams building multi-agent enterprise applications who need a single platform from model fine-tuning through production governance
- Pricing
- Pay-per-use, no monthly fee. Gemini model tokens from $0.10/M input (Flash-Lite) to $1.25/M input (Pro). Agent compute at $0.085/vCPU-hour (50 vCPU-hours free/month). Vertex AI Search at ~$4/1,000 queries. Custom training billed per node-hour by machine type and GPU.
Key features
- 200+ models in Model Garden, including Gemini 3.x, Claude, Gemma, and open-weight models
- Agent Development Kit (ADK) with native A2A protocol support and multi-agent orchestration
- Managed Agent Engine runtime with persistent Memory Bank across conversations
- RAG Engine with Serverless mode and Cross Corpus Retrieval (public preview)
- Vector Search 2.0 GA, AutoML, and MLOps pipelines under an agent-first hierarchy
Quick answers
- What is Gemini Enterprise Agent Platform?
- Formerly Vertex AI, Gemini Enterprise Agent Platform — rebranded at Google Cloud Next 2026 — is Google Cloud's consolidated runtime for building, deploying, and governing production AI agents. It bundles the Agent Development Kit, Agent Studio (low-code), 200+ foundation models via Model Garden (including Gemini and Claude), a managed agent runtime with persistent Memory Bank, AutoML, and MLOps pipelines under one pay-per-use platform. Existing Vertex AI workloads, SDKs, and APIs continue to run unchanged under the new branding.
- Why does Gemini Enterprise Agent Platform matter?
- The Vertex AI → Gemini Enterprise Agent Platform rebrand is not cosmetic: Google has inverted the product hierarchy so that agents are now the primary organizing concept, with model training and AutoML demoted to sub-menus. This signals Google's bet that enterprise AI spend will shift from standalone model API calls to orchestrated, multi-step agent workflows — and locks GCP customers deeper into a unified billing surface that spans inference, retrieval, memory, and compute. Teams should watch for pricing step-ups as introductory Gemini 3.x Flash rates are scheduled to double in early 2027, and for growing multi-cloud friction as the platform tightens around Google's own A2A and ADK toolchain.
- What is Gemini Enterprise Agent Platform best for?
- GCP-native engineering teams building multi-agent enterprise applications who need a single platform from model fine-tuning through production governance
- How much does Gemini Enterprise Agent Platform cost?
- Pay-per-use, no monthly fee. Gemini model tokens from $0.10/M input (Flash-Lite) to $1.25/M input (Pro). Agent compute at $0.085/vCPU-hour (50 vCPU-hours free/month). Vertex AI Search at ~$4/1,000 queries. Custom training billed per node-hour by machine type and GPU.
Seen in recent stories
Where Gemini Enterprise Agent Platform showed up in the last 90 days of KeyNews editions.
Know the terms
Glossary entries that explain what Gemini Enterprise Agent Platform is built on.
Agentic orchestration
The architecture pattern of coordinating multiple AI agents to accomplish complex tasks, with a supervisor agent routing work, managing state, and combining results from specialized sub-agents.
Agentic RAG
A retrieval-augmented generation system where an AI agent autonomously decides when, what, and how to retrieve information — dynamically choosing between multiple knowledge sources, reformulating queries, and iterating on results rather than making a single fixed retrieval call.
Hyperscaler
A cloud computing provider operating at massive scale, primarily Microsoft Azure, Amazon AWS, and Google Cloud. Hyperscalers provide the GPU infrastructure, managed AI services, and global data center networks that power most AI deployments.
A2A (Agent-to-Agent)
A protocol that enables AI agents built by different vendors to discover, authenticate, and collaborate with each other. A2A standardizes how agents delegate sub-tasks, share context, and return results across organizational boundaries.
Models behind it
The models Gemini Enterprise Agent Platform runs on, offers, or competes with.
New tools drop every week.
ONE BRIEFING · EVERY FRIDAY · FREE
Free. Unsubscribe anytime.