Google Unveils Gemini Enterprise Agent Platform, Consolidates AI Stack at Cloud Next 2026
Key Points
- Google renames Vertex AI to Gemini Enterprise Agent Platform and folds Agentspace into a unified product.
- Workspace Studio launches as a no‑code builder for AI agents across Gmail, Docs, Sheets, Drive, Meet and Chat.
- Model Garden now hosts more than 200 models, including Google’s Gemini, Anthropic’s Claude and open‑source Llama.
- Project Mariner, a web‑browsing agent, scores 83.5% on the WebVoyager benchmark and is available to AI Ultra subscribers.
- Agent2Agent protocol reaches production use at 150 enterprises, enabling cross‑platform agent communication.
- Apigee becomes a bridge for Model Context Protocol, turning standard APIs into discoverable agent tools.
- Security upgrades include Model Armor, zero‑trust architecture and cryptographic signatures for domain verification.
- Early customers report up to 95% reduction in query time and near‑real‑time email order processing.
At Cloud Next 2026, Google announced a sweeping rebrand of its AI offerings, merging Vertex AI, Agentspace and new services into the Gemini Enterprise Agent Platform. The rollout includes Workspace Studio, a no‑code builder for AI agents across Gmail, Docs and other Google apps, a Model Garden with more than 200 models, and the production‑grade Agent2Agent protocol now in use at 150 enterprises. CEO Thomas Kurian framed the move as owning the entire stack—from custom TPU silicon to the inbox—positioning Google against rivals that, he said, “hand you the pieces, not the platform.”
Google used the opening keynote of Cloud Next 2026 to announce a full rebranding and consolidation of its AI platform around agents. Vertex AI is now the Gemini Enterprise Agent Platform, and the employee‑facing assistant formerly known as Agentspace has been folded into a unified Gemini Enterprise product. The shift reflects a strategic push to control every layer of the AI stack—from custom Ironwood TPUs to the productivity tools that millions use daily.
Thomas Kurian, head of Google Cloud, framed the strategy as “owning the full stack from chip to inbox.” He contrasted Google’s integrated approach with competitors that, in his words, “hand you the pieces, not the platform.” The announcement bundles several new capabilities: a no‑code builder called Workspace Studio, an expanded Model Garden with over 200 models, a web‑browsing agent named Project Mariner, managed MCP servers, and a production‑grade Agent2Agent (A2A) protocol now deployed in 150 organizations.
Workspace Studio makes AI agents accessible to business users. The new tool lets users create and deploy agents across Gmail, Docs, Sheets, Drive, Meet and Chat by describing tasks in plain language. A simple prompt such as “every Friday, ping me to update my tracker” generates a fully functional automation. The builder connects to third‑party apps like Asana, Jira, Mailchimp and Salesforce, and can invoke external APIs via webhooks or custom Apps Script logic. Google is rolling the feature out to Workspace business, enterprise and education customers.
On the developer side, the Gemini Enterprise Agent Platform now includes a visual flow canvas called Agent Designer, persistent context features (Agent Engine Sessions and Memory Bank), and a pre‑built Agent Garden for customer service, data analysis and creative tasks. The Model Garden hosts Google’s own Gemini and Gemma families, third‑party models such as Anthropic’s Claude, and open‑source options like Llama. Google also announced six new agents for BigQuery that automate pipeline creation and translate natural‑language queries into executable Python with visualizations.
Project Mariner, a DeepMind‑powered web‑browsing agent, scored 83.5% on the WebVoyager benchmark and can handle ten concurrent tasks on cloud‑based VMs. It automates shopping, information retrieval and form‑filling, and is currently available to Google AI Ultra subscribers in the United States. A visual builder for Mariner is slated for the second quarter, with cross‑device sync and an agent marketplace planned for later in the year.
The most strategically significant, yet least visible, announcement was the Agent2Agent (A2A) protocol. Originally launched with more than 50 technology partners, A2A now routes real tasks between agents built on different platforms in production at 150 enterprises, including Microsoft, AWS, Salesforce, SAP and ServiceNow. Governed by the Linux Foundation’s Agentic AI Foundation, the protocol complements Anthropic’s Model Context Protocol (MCP) by handling inter‑agent communication across organizational boundaries. Google’s Apigee now serves as an MCP bridge, turning any standard API into a discoverable agent tool with existing security and governance controls.
Security and governance received a boost as well. The Agent Development Kit reached stable v1.0 across Python, Go and Java, added TypeScript support, and incorporated Model Armor to defend against indirect prompt injection. Zero‑trust architecture, Cloud IAM audit logging and cryptographic signatures for domain verification underpin the new ecosystem.
Google’s move arrives as rivals double down on their own agent strategies. OpenAI’s Operator, Anthropic’s Claude marketplace, Microsoft’s Copilot and AWS’s Bedrock are all gaining traction. Google, however, argues that its vertical integration—custom silicon, frontier models, a unified cloud platform and the massive Workspace user base—gives it a unique advantage. An internal AI Agent Trends report found that 89% of business teams already use AI agents, with an average of 12 agents per organization. Early adopters like Danfoss and Suzano report dramatic productivity gains, cutting email‑order processing times from 42 hours to near real‑time and reducing query times by 95% for tens of thousands of employees.
While Google Cloud’s market share remains behind AWS (31%) and Azure (25%), the company posted the fastest year‑over‑year growth among the three cloud providers in Q4 2025, up 50%. The Gemini Enterprise Agent Platform is the centerpiece of that growth, aiming to make the “agentic cloud” the default productivity layer for enterprises worldwide.