Table of Content
AI Information
AI Tutorial
AI Agent
Latest Information
Hot Information
LLMs Information
Technical Frontiers
Industry Applications
Prompt Tutorial
AI Programming
AI Video Generation
AI Audio Generation
Knowledge Base Setup
AI Development Guides
Industry Tutorials
Practical Cases
AI Resource Library
Agent Information
Agent Setup
Agent Training

Cloud Manufacturers Rush to MCP Track: Key Turning Point for AI from Chatbots to Digital Partners

Written by
Silas Grey
Updated on:June-10th-2025
Recommendation

AI assistants are evolving from chat tools to all-around digital partners, and the MCP protocol is the key to this change. Core content: 1. Definition of MCP protocol and expansion of AI capabilities 2. Construction of the MCP service ecosystem and 10 major categories of applications 3. How cloud vendors plan competition in the MCP field

 
Yang Fangxian
53A founder/Tencent Cloud (TVP), most valuable expert
Did you notice that the recent AI assistants are no longer limited to chatting? They can check tickets, analyze stocks, summarize documents, and even operate design software. 
Behind this, a protocol called MCP (Model Context Protocol) and its ecosystem are quietly changing the boundaries of AI's capabilities.

01

 

 

 

 

What is MCP?

Imagine a knowledgeable scholar (LLM) who is full of knowledge but cannot directly access external information or operating tools. The emergence of the MCP protocol is like opening doors and windows for this scholar, and equipped with a capable assistant. Its core goal is to provide standardized interfaces for AI models (LLMs) that enable it to call external tools, services, or data sources safely and reliably.

(For the past and present of MCP, you can read Xiaomeng's previous articles, Agent era entrance: From competing for the standard for large language model plug-in to embracing MCP, everyone is rushing to the entrance to the Agent era .)

Now, the entire MCP industry chain is divided into three parts:

  • It is those AI chat interfaces, Claude, ChatGPT, etc.

  • MCP Server: Various tool services, search, map, file processing, etc.

  • MCP Marketplace: Tool store, similar to App Store

02

MCP Service = AI Toolbox

MCP services have built a complete ecosystem covering multiple fields, which can be roughly divided into 10 categories: developer tools, search tools, knowledge management and memory, communication and collaboration tools, finance, academic research, file systems, entertainment and multimedia, location services, schedule management, etc.

Covers all aspects from programming development, information retrieval, office collaboration, to life and entertainment, providing AI assistants with rich external capabilities expansion. The following trends are presented:

  • Technologically driven ecological pattern: Developer tools and data processing tools occupy a core position

  • Platform integration becomes a key capability: The core value of MCP services lies in breaking through data barriers between different platforms and realizing cross-platform operation capabilities of AI

  • Real-time information acquisition and web data processing have become important directions for the expansion of AI capabilities

  • Rapid growth of enterprise-level applications: Communication and collaboration tools focus on enterprise scenarios

( Here Xiao Meng has compiled a directory called "Best-use MCP Server". If you need a private message to the background, you can directly get it~ )

03

 

 

 

 

Major cloud manufacturers place crazy bets


MCP is still in its early stages, and the pattern is far from settled. Cloud manufacturers such as Alibaba, Tencent, and Baidu have been crazily including MCP services recently to see who can build the largest ecosystem.

How do these mainstream manufacturers plan the MCP market?

Technical architecture: Full-stack hosting, lowering the threshold for enterprise-level applications

  • Full life cycle MCP service: All manufacturers provide the ability to "fast build an Agent in 5 minutes". Alibaba Cloud Bailian integrates function calculation + 200+ models + 50+ tool chain, Tencent Cloud realizes MCP plug-in hosting through AI development kit, Baidu Qianfan supports SDK to convert MCP Server in one click, and no user management is required.

  • Performance and security optimization: low-latency processing, enterprise-level security

Ecological strategy: closed-loop ecological binding, strengthen B-end penetration

  • Ali: Convert 50+ applications such as Gaode and Cainiao into MCP infrastructure to form a closed loop of "applications are protocols".

  • Tencent: Open up the WeChat/enterprise WeChat portal and reach 1 billion users through the MCP Marketplace.

  • Baidu: Integrate applications such as Baidu Library, Baidu Netdisk, Baidu Maps

Monetization mode: Cloud consumption base + ecological distribution amplification

  • Cloud resource consumption share

  • MCP service transaction commission

  • MCP service transaction commission

  • MCP service transaction commission

  • Solution

  • Trade sharing

As the standardization of the MCP protocol advances, manufacturers need to find a balance between openness and control, and compete for the ecological dominance of the "first year of AI intelligent body."

When the boundaries of AI's capabilities are infinitely extended through the MCP protocol, will our relationship with artificial intelligence also undergo a fundamental reshaping? It is no longer just a "wise man" who answers questions, but will also become a "digital partner" who can work on behalf of others, collaborate, and even actively plan.

This is no longer a technical question about "what can AI do", but a future proposition about "how we can coexist with stronger AI." Understanding MCP is the key to understanding the opening of this new paradigm of human-computer collaboration.

 

Related PostsView More
Beyond the Basics: How Agentic Chunking Revolutionized RAG
In-depth discussion of the changes in the field of retrieval-augmented generation (RAG), focusing on how Agentic Chunking completely subverts tradition. Understand the limitations of current mainstream chunking methods such as fixed size and recursion, as well as the core idea of ​​Agentic Chunking to improve text chunking with intelligent decision-making. It uses LLM's understanding,...
Caleb Hayes
Re-understanding the boundaries and potential of Agents | AI Transformation Interview
In-depth discussion of the boundaries and potential of large language models in the context of 2025, the "first year of Agents"! Dr. Fan Ling, founder of Tezan, shared Atypica.ai's unique insights, breaking through traditional frameworks, simulating real user behaviors, and multiple AI assistants working together to conduct large-scale user interviews efficiently and at low cost. Its divergent...
Iris Vance
A simple framework for solving the RAG question-answering hallucination problem based on KG generated corpus enhancement - Walk & Retrieve
In-depth exploration of a simple framework based on KG generated corpus enhancement to solve the RAG question-answering hallucination problem - Walk&Retrieve! This framework uses knowledge graphs to generate zero-shot RAG corpora, effectively solving the RAG system hallucination. It includes graph-based traversal methods such as random walks and BFS walks, as well as key steps such as...
Clara Bennett