Google has announced plans to integrate Anthropic’s Model Context Protocol (MCP) into its Gemini models and software development kits (SDKs), marking a significant shift in how AI models will connect to external data systems. The move comes just weeks after OpenAI revealed it would follow a similar approach, embracing MCP as a standard for AI systems to seamlessly access and interact with data.
In a post on X (formerly Twitter), Demis Hassabis, CEO of Google DeepMind, expressed his support for MCP, calling it a “good protocol” and a rapidly emerging “open standard for the AI agentic era.” He added, “Look forward to developing it further with the MCP team and others in the industry.”
The Model Context Protocol developed at Anthropic allows AI models to access different data sources including business tools and software as well as content repositories and application environments. The integration power of MCP allows developers to make sustainable data connections between AI systems and repositories which results in improved applications like chatbots and AI-powered workflows.
Developers who support MCP enable “MCP servers” data exposure for applications which they build using “MCP clients” to connect dynamically to MCP servers. MCP helps AI models conduct seamless data exchanges with external systems thus resulting in faster completion of tasks and improved performance for AI-powered tools.
Since Anthropic open-sourced MCP multiple businesses have integrated the protocol across their operations including Block, Apollo, Replit, Codeium and Sourcegraph. The industry is moving toward AI data standardization with Google’s support for MCP suggesting future developments of scalable robust AI applications throughout different industries.
While Google did not provide a specific timeline for when MCP will be available in its Gemini models and SDKs, the company’s announcement underscores the growing importance of open standards in AI development, with leading tech companies like Google and OpenAI driving the future of AI interoperability and integration.
