The news, shared on April 10 via Google DeepMind CEO Demis Hassabis in a post on X (formerly Twitter), comes just two weeks after OpenAI announced its own adoption of MCP. With two of the world’s most influential AI companies now backing it, MCP is fast becoming the go-to standard for how AI models plug into real-world workflows.

In what might be a rare show of unity in the AI race, Google is following OpenAI’s footsteps and embracing a shared tech protocol — something that’s not very common among rival AI giants. The company has announced that it will support Anthropic’s Model Context Protocol (MCP), a tool that makes it easier for AI systems to connect with business apps and data.

What is MCP and why does it matter?

Think of MCP like a universal power adapter — but for AI agents. Instead of building a separate setup to help an AI model work with Google Drive, Slack, or GitHub, developers can use MCP to connect their model to all of them in a single shot. The protocol helps AI systems get the context they need by tapping into different apps and tools seamlessly. AI agents — the digital workers that are increasingly showing up in our apps and workflows — are only as useful as the data they can access. Without deep integration with tools and services, they’re more like clever chatbots than reliable assistants. That’s where MCP comes in. “MCP is a good protocol, and it’s rapidly becoming an open standard for the AI agentic era,” said Hassabis. “We’re excited to announce that we’ll be supporting it for our Gemini models and SDK.”

Open source and industry-wide support

Anthropic, the startup behind Claude AI, open-sourced MCP in 2024. Since then, several companies have joined the effort, including Replit, Codeium, Apollo, Zed, Block, and Sourcegraph. Anthropic even shared pre-built servers for popular software platforms so developers could get started right away. These include servers for tools like GitHub and Google Drive, making it easier for companies to hook everything up with less hassle.

Earlier, OpenAI CEO Sam Altman said MCP would roll out first in OpenAI’s SDK and eventually to ChatGPT on both desktop and mobile. Google hasn’t provided an exact timeline for its Gemini models and SDK to support MCP yet.

What’s next?

For now, the industry seems to agree on one thing — making AI agents more useful means giving them better access to the tools and data we already use. And MCP could be the plug that powers all of it.

The move signals a possible shift in the way AI development is heading — one where even major competitors can agree on common tools, especially as AI agents grow more complex. Whether this leads to more open collaboration or simply more efficient products remains to be seen.

Read More:

No code? No problem! Google’s Firebase Studio helps build AI apps in your browser