Representative End User Clients
Representative Automation Clients
Representative Software Clients
As we continue to explore Anthropic’s Model Context Protocol (MCP) and its implications for Industrial AI, a key question from the industry has surfaced: Does MCP replace existing standards like OPC UA, or is it designed to complement them? This is an important distinction, especially considering OPC UA’s critical role in today’s industrial connectivity. Having previously discussed MCP’s potential to standardize AI context in my previous blog (and the ones preceding it), it’s worth clarifying how it relates to established frameworks—drawing on ARC’s ongoing analysis.
Our findings confirm a clear answer: MCP complements OPC UA; it does not replace it. These technologies serve distinct, vital functions within the modern Industrial Data Fabric architecture needed for effective Industrial AI.
OPC UA: The Bedrock of Industrial Data Acquisition
First, it’s important to reaffirm the role of OPC UA. It serves as a secure, platform-independent, and widely adopted standard for industrial communication and data acquisition. Its strength lies in connecting a broad spectrum of industrial assets—PLCs, sensors, SCADA systems, and more—while standardizing data exchange at the operational technology (OT) layer. OPC UA provides:
Broad Connectivity: It bridges the gap between diverse industrial hardware and software.
Standardized Data Models: Through its information models and Companion Specifications, it ensures data is not just transmitted but also understood semantically.
Security and Reliability: It offers robust mechanisms tailored for industrial environments.
Essentially, OPC UA is the workhorse ensuring reliable, standardized data flows from the plant floor into the broader data infrastructure, like an Industrial Data Fabric.
MCP: Standardizing Context for AI
Anthropic's MCP, as discussed earlier, addresses a distinct challenge. Its main goal is to standardize how applications deliver contextual information to AI models, especially Large Language Models (LLMs). MCP functions as a universal interface, enabling AI to securely access and interact with external data sources and tools beyond its training data. It focuses on:
Context Provisioning: Enabling AI to ground its responses in relevant, real-time, or specific external data.
Tool Integration: Enabling AI models to invoke external functions or APIs through a standardized interface.
Simplifying AI Integration: Addressing the "N×M" problem of connecting numerous AI models to numerous tools/data sources.
MCP functions at the intersection of the AI model and the data or tools it relies on, ensuring the model accurately interprets the context of the information it receives and the actions it is permitted to perform.
Why MCP does not Replace OPC UA
A closer look reveals key distinctions that prevent MCP from being a substitute for OPC UA:
Different Focus: OPC UA is fundamentally about device-level communication and data acquisition. MCP is about AI model context and tool interaction.
Hardware Interaction: MCP is not designed to directly interface with the wide range of industrial hardware and protocols that OPC UA supports natively. OPC UA operates deep within the automation pyramid, often connecting directly to field devices.
Established Ecosystem: OPC UA benefits from a deeply entrenched and mature ecosystem within industrial automation, representing significant existing infrastructure.
MCP requires access to data, and in industrial environments, OPC UA is frequently the primary and trusted method for reliably collecting and initially standardizing critical operational data.
Synergy in the Industrial Data Fabric
The real power emerges when these technologies work together within an Industrial Data Fabric, often orchestrated by a Unified Namespace (UNS):
Data Acquisition (OPC UA): OPC UA collects raw and structured data from industrial assets.
Organization and Contextualization (UNS): This data flows into the Industrial Data Fabric, where a Unified Namespace (UNS) structures it into a logical, semantic hierarchy (often via MQTT), adding business context and serving as the “single source of truth”.
AI Access and Understanding (MCP): MCP servers can then connect with the UNS, enabling AI models (through MCP clients) to securely query and access specific, contextualized data required for tasks such as predictive maintenance, process optimization, or quality analysis. MCP ensures this structured data is delivered in a standardized format that the AI can readily interpret and utilize.
The Path Forward: Collaboration is Key
Unlocking the full potential of this synergistic approach goes beyond understanding each technology in isolation. It calls for increased industry collaboration to establish best practices and, potentially, standardize how these layers are integrated. This is where standards bodies and industry organizations focused on smart manufacturing, like CESMII (the Smart Manufacturing Institute), which we highlighted in our third blog, can play a crucial role. Providing clear guidance on how OPC UA, UNS architectures, and MCP can seamlessly interoperate is essential for building robust, scalable, and secure industrial-grade data fabrics that support advanced AI applications. Cultivating an ecosystem where connectors and integration patterns are openly shared will help accelerate both adoption and innovation.
Building on Existing Foundations
So, rather than replacing OPC UA, MCP builds upon the foundation it provides. OPC UA secures the vital data pipeline from the OT layer, the UNS structures and contextualizes this data within the Industrial Data Fabric, and MCP provides a standardized way for Industrial AI models to leverage this rich, contextualized information effectively and securely.
Grasping this synergy is crucial. For organizations advancing Industrial AI, the goal isn’t to choose between OPC UA and MCP, but to integrate them effectively—using OPC UA for reliable data acquisition and MCP to streamline how AI models access and apply that data within a well-designed Industrial Data Fabric. Promoting industry collaboration on integration standards will strengthen this approach, enabling more dependable and context-aware Industrial AI solutions.
Links to Previous Blogs on the Model Context Protocol
Beyond the Hype: Practical Realities of MCP in Industrial AI Deployment
The Power Trio: MCP, UNS, and CESMII–Orchestrating the Future of Industrial AI
Contextual Harmony: Industrial AI, Unified Namespaces and the Promise of MCP
Context Is the Missing Link: The Emergence of the Model Context Protocol in Industrial AI
Engage with ARC Advisory Group
For ARC Advisory Group recommendations for navigating the AI Wars, closing the digital divide by embracing Industrial AI, assembling your Industrial-grade Data Fabric, and governing and guiding major decisions about enterprise, cloud, industrial edge, and AI software, please contact Colin Masson at [email protected] or set up a meeting with me, or my fellow Analysts at ARC Advisory Group.