Prime Highlight

• Google introduced fully managed MCP servers that allow AI agents to easily connect with its tools and cloud services, reducing setup time and integration complexity.
• The servers enable agents to access live data from services like Maps, BigQuery, and Compute Engine, ensuring responses are grounded in up-to-date information.

Key Facts

• MCP, created by Anthropic and part of a Linux Foundation fund, is an open standard compatible with multiple AI clients, including Gemini, Claude, and ChatGPT.
• The servers are currently in public preview at no extra cost for enterprise customers, with general availability expected early next year.

Background

Google launched a new line of fully managed MCP servers to help AI agents connect easily with its tools and cloud services. The new servers aim to solve a common problem for developers, who often find it hard to link AI systems with real-world data using unreliable custom connectors.

The new servers allow agents to plug into Google services such as Maps, BigQuery, Compute Engine, and Kubernetes Engine through a managed endpoint. The company says this will cut down weeks of setup work and give agents more reliable access to live data.

“We are designing our products to be Google agent-ready from the start,” said Steren Giannini, director of product management at Google Cloud. He added that developers can now simply paste a URL instead of building and maintaining multiple integrations.

With this new setup, a travel-planning agent can access real-time location data from Maps, while an analytics agent can query BigQuery directly. Giannini noted that, without MCP, developers rely on a model’s built-in knowledge, which may not always be current. The MCP server removes that gap by grounding responses in updated information.

The launch follows Google’s latest Gemini 3 model, as the company tries to combine improved reasoning with dependable access to tools. The servers are currently available in public preview and come at no extra charge for current enterprise customers. General availability is expected early next year.

MCP, created by Anthropic, is an open standard and now part of a Linux Foundation fund aimed at building shared infrastructure for AI agents. Because it is a standard, Google’s MCP servers can work with multiple clients, including Gemini, Claude, and ChatGPT.

Google says the larger enterprise benefit comes from Apigee, which can convert existing APIs into MCP servers, adding governance, security controls, and audit logs for AI agents. The company intends to extend support to additional Google Cloud services in the coming months.

Giannini explained that developers no longer need to handle the plumbing themselves because “we built it for them.”