For the past two years, the public conversation about artificial intelligence has been dominated by model launches, benchmark scores, and leaderboard bragging rights. That framing is increasingly misleading. The most important strategic moves in AI are not happening at the model layer at all. They are happening one layer down, in the protocols, connectors, and integration standards that determine how models reach real work.

That is why the recent wave of protocol giveaways matters. Anthropic donated MCP, the Model Context Protocol, to the Linux Foundation’s new Agentic AI Foundation. At the time, MCP was pulling nearly 100 million monthly SDK downloads across more than 10,000 active servers. Google handed its Agent2Agent protocol, A2A, to the same foundation, with AWS, Cisco, Microsoft, Salesforce, SAP, and ServiceNow signing on as founding members. OpenAI supports remote MCP servers in its Responses API, sits on the MCP steering committee, and contributed AGENTS.md to the foundation alongside its fiercest rival’s protocol.

On the surface, these look like acts of generosity. They are better understood as competitive strategy. Platform companies do not usually give away technology they still expect to be the primary source of advantage. They give it away when the advantage has moved somewhere else.

Key facts at a glance

• Anthropic transferred MCP to the Linux Foundation’s new Agentic AI Foundation after the protocol reached nearly 100 million monthly SDK downloads and more than 10,000 active servers.
• Google donated A2A to the same foundation, with AWS, Cisco, Microsoft, Salesforce, SAP, and ServiceNow as founding members.
• OpenAI supports remote MCP servers in its Responses API, participates in MCP steering, and contributed AGENTS.md to the foundation.
• The strategic logic is not charity but commoditization: reduce friction and lock-in at one layer so competition can shift to another.
• Durable AI advantage is moving toward enterprise data, workflows, permissions, governance, and the places where AI-assisted work accumulates.

The protocol giveaway is not charity

The pattern is familiar to anyone who has watched platform markets mature. Companies routinely commoditize one layer of the stack so they can compete somewhere they hold a stronger hand. The donated protocol is not the product. It is the on-ramp. It is the road that makes it easier for developers and enterprises to connect models to tools, databases, applications, and agents. Once that road is open, the fight moves to the destination.

That is the key distinction. Open protocols lower the cost of connection, but they do not automatically lower the cost of switching the systems where work actually lives. A developer may be able to swap one MCP-compatible model for another. An enterprise may be able to adopt a new agent framework that follows common standards. But the prompts, evaluations, security policies, approval workflows, employee habits, and integration history do not move as easily. Advantage accumulates around those things.

Why the model layer is a weak moat

The frontier model itself has turned out to be a poor place to build a lasting moat. Leaderboards change almost weekly. A model that looks dominant today can be matched or surpassed tomorrow. Enterprises should not build their AI strategy around the assumption that any single vendor will remain permanently ahead on model quality. The dull reality is that connecting models to enterprise data, workflows, and governance is where the durable value is created.

The major AI companies understand this better than their public messaging sometimes suggests. They will keep spending billions to train more capable models because frontier models attract developers, generate headlines, and open enterprise doors. But they are also quietly acknowledging that benchmark leadership alone does not create a durable platform. A platform needs gravity. It needs developers to return to the same tools, enterprises to connect the same systems, and workflows to accumulate around the same set of interfaces.

The cloud and open-source playbook

Google offered an earlier version of this playbook. The company has long been an exceptionally active contributor to open source. It did not open source TensorFlow and Kubernetes out of pure generosity. Those projects became on-ramps for Google Cloud at a time when Google was playing catch-up to AWS and Microsoft. The hope was to create a gravity well in the market for container-based applications so that a significant percentage of them would end up on Google’s platform.

GitHub may be an even better example. Git is open. Anyone can host a Git repository, and once upon a time different companies did just that. Yet GitHub became the default place where software development happens for millions of developers. Nobody pays for Git, but lots of people pay for GitHub. The open protocol did not eliminate competition. It moved competition to the place where collaboration, workflows, and accumulated project history created stickiness.

Developers return to gravity wells

AI is following the same path. Developers return to the places where their tools, workflows, teammates, and accumulated work already live. Enterprises double down on the systems where their data, permissions, governance, and business processes are already connected. Every new integration makes that destination a little harder to leave. Every new workflow increases its pull. That is what MCP, A2A, and similar efforts are really about: increasing gravity around the models.

Every major AI company wants to become the place where AI-assisted work naturally happens. They are amassing armies of forward-deployed engineers and using other means to tie legacy infrastructure back to their frontier models. Enterprise incumbents want the same thing from the opposite direction. They do not need to own the frontier model. They need to connect the frontier to the systems that already run the business safely and reliably.

Enterprises do not live on the frontier

That is why predictions that AI will simply sweep away enterprise software deserve skepticism. Developers absolutely live on the frontier. Enterprises do not. Enterprises create value by connecting new capabilities to decades of accumulated applications, data, policies, and business processes. The newest model matters. The newest agent framework matters. But neither creates much business value until it is connected to customer records, financial systems, supply chains, HR data, and everything else enterprises already depend on.

That is where incumbents still possess enormous gravitational pull. Large enterprise software vendors understand that their advantage is not a single model. It is the installed base of systems, the integration surface, the governance controls, and the institutional memory embedded in their platforms. Ironically, open protocols can strengthen that position rather than weaken it. If every model can speak MCP and every agent can interoperate through common standards, enterprises gain the freedom to adopt whichever frontier technology looks best without rebuilding every integration. The protocol becomes interchangeable. The destination does not.

Open standards move advantage, they do not erase it

None of this diminishes the importance of open standards. MCP succeeded because it solves a genuine problem. Developers should not have to build a custom connector every time an AI application needs access to a database or another business system. Neutral governance also matters because nobody wants foundational infrastructure controlled by a direct competitor. But open interfaces should not be confused with open markets.

An enterprise may find it easy to swap one MCP-compatible model for another while still remaining deeply dependent on the place where its prompts, evaluations, security policies, and employee habits have accumulated. Kubernetes made workloads dramatically more portable without making AWS, Microsoft Azure, and Google Cloud interchangeable. SQL has been standardized for decades, yet databases remain fiercely differentiated businesses. Standards reduce friction, but they rarely eliminate competitive advantage. They simply move it.

In the same way, Anthropic, Google, OpenAI, and others are happily standardizing how models, agents, tools, and enterprise systems communicate because they do not expect the connection itself to determine the winner. They expect to win by becoming the place where AI-assisted work naturally accumulates. Along the way, the market will see copious quantities of code given away, rising developer productivity for many, and outsized financial gains for a few.

The competition is not over who builds the best model for a single quarter. It is over who becomes the default destination for the work that models make possible. The protocols are open because the real battle is elsewhere: in the enterprise systems, developer habits, and integration gravity that determine where AI-assisted work actually lands.

Source:InfoWorld News