MCP in 2026: 97 Million Downloads and Growing Crypto Infrastructure From Bitgo to Coingecko

The Model Context Protocol (MCP) has recorded approximately 97 million monthly SDK downloads as of March 2026, confirming its position as the primary infrastructure layer for agentic artificial intelligence (AI) applications. Anthropic open-sourced MCP in November 2024, and OpenAI added full support in March 2025, with Google, Microsoft, and others following. In December 2025, Anthropic donated MCP to the Agentic AI Foundation under the Linux Foundation, enhancing its vendor-neutral and community-driven governance.

Mastercard is enhancing its digital currency strategy by launching the Crypto Partner Program and acquiring stablecoin payments provider BVNK for $1.8 billion. This move aligns with Mastercard's focus on enterprise use cases like cross-border payments and faster settlement. The Crypto Partner Program involves 85 crypto-native companies, signaling a shift toward structured, regulated crypto integration.

The crypto x payments sector is progressing beyond experimentation into a phase defined by operational readiness and compliance architecture. Mastercard's Crypto Partner Program is part of a broader trend toward institutional coordination and regulatory adaptability. The Bank of England has signaled openness to revising proposed stablecoin regulations, with potential adjustments expected in June 2026.

What Role Is MCP Playing In The Crypto Sector?

Bitgo launched the Model Context Protocol (MCP) Server to enable AI tools to access and interact with its developer documentation using natural language. This tool aims to streamline development by allowing AI agents to retrieve technical information, generate code snippets, and explain API functionality, reducing the need for manual navigation through documentation. The current version of the MCP Server is limited to documentation access and does not support transaction execution or custody operations.

The BitGo MCP Server supports AI clients in retrieving documentation context while building on BitGo's platform. Developers can now use natural language to explore wallet functionality, transaction flows, staking documentation, webhook configuration, and policy features. This tool is available with compatible AI platforms and aims to streamline the integration of AI into software development.

How Is MCP Shaping AI-Native Infrastructure Across Industries?

Mplify has launched the Kylie SDK release, introducing Model Context Protocol (MCP) support across its Lifecycle Service Orchestration (LSO) APIs. This update enables service providers to automate, monetize, and scale digital services across interconnected ecosystems. The integration strengthens Mplify's role in translating standardized LSO APIs into AI-ready automation frameworks.

The key features of the Kylie SDK include updated packages across Interface Reference Points (IRPs), ensuring support for business and operational functions. AI-Native LSO APIs now support MCP, enabling integration with AI-driven orchestration and diagnostic tools. Enhanced LSO Business and Operational APIs streamline processes like ordering, quoting, billing, service provisioning, and performance monitoring.

The LSO Marketplace serves as the central access point for these capabilities, providing a unified resource for MCP-to-LSO conversion. A major enhancement in the release is the LSO API Blending Tool, which allows users to combine LSO APIs with specific product payloads, such as IP, Carrier Ethernet, wavelengths, and SD-WAN. The Blending Tool can generate Mplify MCP servers with full payload support, enabling seamless conversion from MCP commands to LSO APIs.

The Model Context Protocol (MCP) is an open standard that enables AI models to connect to external tools, databases, and APIs using a universal protocol. Introduced by Anthropic in 2024, it has been widely adopted by major platforms, including OpenAI, Google, and Microsoft. MCP eliminates the need for custom integrations by allowing a single server to serve all compatible clients.

MCP works by defining three key roles: the host (the AI application), the client (which manages connections to servers), and the server (which exposes capabilities to the AI). When a user makes a request, the AI model decides which tools to invoke, and the appropriate client routes the request to the correct server. The server executes the action and returns the result, which the model then incorporates into its response.

MCP supports two transport mechanisms: stdio for local connections and Streamable HTTP for remote servers. The protocol has evolved to include features like async tasks, sampling, and elicitation, enabling more complex workflows. The authentication and authorization mechanisms have matured, including OAuth 2.1 and Resource Indicators to prevent token misuse. The 2026 roadmap focuses on enterprise readiness, transport evolution, and governance maturation.