Leading cloud providers have swiftly embraced AI agents and the Model Context Protocol. Discover how to leverage official MCP servers from major cloud giants to streamline your cloud operations.
Anthropic’s innovative Model Context Protocol (MCP), often dubbed the “USB-C for AI,” has ignited fresh thinking in the software industry regarding AI assistant capabilities. Now, with access to external data, APIs, and internal platforms/databases, AI agents are gaining powerful new abilities for extensive automation.
MCP is no longer exclusive to innovative AI startups or niche SaaS providers. Major cloud platforms are now integrating MCP servers into their offerings, empowering customers to automate crucial cloud computing tasks. These MCP servers act as an AI consumption protocol, complementing existing CLIs and APIs.
By connecting an MCP server to their chosen cloud environment, engineers can significantly enhance the quality of responses from AI coding agents. This is achieved by feeding them cloud documentation and other highly relevant contextual information. Beyond knowledge retrieval, cloud-based MCP can facilitate spinning up new servers, adjusting configurations, or instantly pulling production metadata.
The best part is that any MCP-compliant AI client can seamlessly interact with these servers using simple natural language commands. Popular AI-powered development environments like Cursor, Windsurf, and Visual Studio Code, along with LLM-driven AI agents such as Claude, Codex, and GitHub Copilot, offer out-of-the-box MCP support.
Below, we will explore the MCP servers provided by the leading cloud providers. These officially supported remote MCP servers leverage your existing cloud credentials, enabling authenticated API calls directly from AI clients. They are available for use without additional licensing, incurring only standard cloud service and data transfer costs.
AWS MCP Servers
Amazon Web Services (AWS) offers an extensive collection of over 60 official MCP servers, covering its vast product portfolio. AWS MCP servers address a wide range of functions—from documentation access to infrastructure deployment, container management, Lambda functions, AI/ML frameworks, data and analytics, messaging, cost analysis, and more.
The AWS MCP Server, designed for general purposes, is an excellent starting point. This remote server, hosted by AWS, connects agents with the latest documentation, API references, and standard operating procedures (SOPs) to execute complex, multi-step workflows. These SOPs can handle infrastructure provisioning and configuration, as well as cloud cost monitoring and analysis.
Consider a real-world MCP scenario: troubleshooting errors affecting multiple AWS services. You could query the AWS MCP Server with a command like, “Investigate the increase in 5xx errors in production over the last 30 minutes.” Given the appropriate context and permissions, the AWS MCP Server would then access relevant metrics, logs, and configuration data across services to pinpoint a probable root cause.
A significant advantage of AWS’s MCP strategy is the official maintenance of these servers. The comprehensive catalog continues to grow and is gradually transitioning to Streamable HTTP, an enhanced transport protocol. Clearly, AWS has made substantial investments in MCP, establishing it as a core component for agent-driven cloud operations.
Azure MCP Server
Microsoft Azure’s Azure MCP Server empowers AI agents to interact with Azure services using natural language commands. Rather than offering separate MCP servers for each service, Azure structures its MCP server into more than 40 individual MCP tools, covering Azure best practices, AI/ML services, analytics, compute, containers, databases, DevOps, IoT, storage, and other essential categories.
With the Azure MCP Server, you can engage conversationally with Azure, using prompts such as “Show me all my resource groups” or “List blobs in my storage container named ‘documents,'” as outlined in the documentation. Other queries allow listing databases, enumerating Azure storage accounts, analyzing large datasets in Azure databases, and executing many additional actions.
Azure provides a straightforward getting-started guide, offering more detailed assistance compared to AWS. The documentation clearly explains installation, tool parameters, and settings for enabling or disabling agent control over sensitive functions. Each tool is well-documented with examples of potential natural language prompts.
Google Cloud MCP Servers
Google Cloud Platform (GCP) unveiled its official Google Cloud MCP servers in December 2025. At present, these remote MCP servers are still in preview, meaning they are offered “as is” with limited support. Despite this, Google Cloud currently provides four operational official remote MCP servers, covering dataset operations, virtual machine management, Kubernetes management, and more.
For example, a natural language query such as “Get metadata details and table IDs for my dataset ‘newUsers'” directed to the BigQuery MCP server would likely trigger tools like list_table_ids to retrieve table IDs, along with get_dataset_info and get_table_info to fetch metadata.
Alternatively, you could issue a command like “Kill my running VM in project 0009 in the east zone” to the Compute Engine MCP. This would invoke the stop_instance tool to halt the VM. Other tools facilitate actions such as deleting or resetting instances, as well as more benign commands like retrieving compute metadata and operational traces.
Google Cloud also offers MCP servers for Google Kubernetes Engine (GKE) and Google Security Operations. Additionally, Google provides the Maps Grounding Lite MCP server, designed to assist developers in building LLM applications on the Google Maps Platform, alongside numerous open-source servers intended for local hosting.
Similar to other cloud offerings, Google Cloud MCP servers include controls to enable either read-only or read-write functions. A unique advantage is Google’s approach to logging for all MCP interactions and access, which greatly aids auditing for cloud administrators. While GCP currently presents a more limited selection of MCP servers compared to other hyperscalers, its MCP tools show significant promise for automating essential cloud computing operations.
Oracle MCP Servers
Oracle, with its established history in providing both private and public cloud solutions for enterprises, has recently ventured into MCP with a concise suite of MCP servers. These servers are designed to interface with popular Oracle platforms, enabling management of Oracle Cloud Infrastructure (OCI) and operations on Oracle databases and MySQL resources.
For instance, Oracle SQLcl, the command-line interface (CLI) for Oracle Database, has an MCP server that allows agents to execute queries and process results. On the Oracle blog, engineers suggest a prompt like, “Connect to my fun side project and tell me about what kind of data I have there,” which would activate a list-connections tool to display all saved Oracle connections.
Other applications for Oracle’s MCP servers include describing and generating database schemas in plain language, real-time analysis of MySQL usage patterns, or linking a project to existing database tables to populate application data.
While some of Oracle’s MCP initiatives are still in the proof-of-concept stage, they indicate a compelling future for integrating established database platforms with nascent AI-driven prototyping and development workflows.
IBM Cloud MCP Servers
The IBM Cloud MCP servers are currently experimental, but they are conceived as a comprehensive knowledge-gathering layer between AI assistants and the IBM Cloud platform. Their primary function is to retrieve information about services within a user’s IBM Cloud computing environments.
Unlike most MCP servers on this list, which operate in the cloud, IBM’s Core MCP Server is designed for local installation, then configured to interface with the IBM Cloud CLI, essentially acting as a layer on top of it. This Core MCP Server can also be containerized for specific use cases. However, there are a few considerations: it is stateful, not built for multi-account use, and lacks OAuth support.
Nonetheless, the Core MCP Server offers a user-friendly method to query IBM Cloud for discovering cloud resources, retrieving extensive metadata, filtering results by strings, listing service names, and more. The documentation provides straightforward prompts such as “Are there any VPCs in this account?”, “What zones are available in us-east?”, and “What resource groups are in my account?”.
Beyond the Core MCP Server, IBM Cloud also offers MCP servers for Cloud Internet Services (covering DNS, GLB, WAF, DDoS, and CDN), logs, streams, Kubernetes and OpenShift, code monitoring, object storage, serverless services, VPC, and other IBM Cloud services.
The documentation is comprehensive, with solid examples, making MCP feel intuitive for those already familiar with IBM’s detailed CLI and API commands. However, most actions available via IBM Cloud MCP servers are read-only. For now, IBM’s MCP servers primarily serve as an experimental interface for information gathering.
The Cloud Awaits Your Command
MCP has seen increasing adoption within enterprises in recent months, a trend amplified by growing hyperscaler support for the protocol. When applied in a cloud operational context, MCP holds the potential to eliminate repetitive tasks, such as configuring fields in graphical user interfaces or manually sifting through API references and product documentation.
Utilizing the MCP servers discussed above presents an exciting opportunity: a novel, streamlined, AI-driven control layer for managing hyperscale cloud environments. This is the ultimate aspiration. In reality, it’s still early days, with many of these servers remaining in experimental or preview stages. Security models also exhibit considerable variation, and not all support mutating operations, frequently defaulting to read-only modes.
Harnessing AI agents, MCP, and natural language for automating cloud operations will demand significant experimentation, practical testing, and creative problem-solving. From database queries and resource management to provisioning, scaling, root-cause analysis, and cost optimization, it’s ultimately up to individual operators to determine how MCP best integrates into their workflows. In essence, with MCP, the cloud is yours to command. What will you create?
