An open standard that any vendor can plug into changes who gets to compete, in a way one company's private API doesn't — why?

This explores why an open, shared protocol changes who gets to compete, while an API owned by one dominant platform tends to lock existing power in place. The core argument in the corpus comes from O'Reilly. Protocols are not neutral plumbing. They shape markets by turning winner-take-all races into webs where many vendors compete Do protocols reshape markets by enabling multi-vendor competition?. When one company owns the API, it sets the rules of access and can change them, price them, or revoke them. When a protocol is shared, nobody holds that switch. The point of control moves away from a single gatekeeper and into a common standard that any vendor can build on. O'Reilly's evidence is MCP: once tools could connect to any AI chatbot through the same interface, which chatbot you used mattered less as a moat.

The less obvious part is that protocols gain market power by fitting in, not by taking over. Research on agent coordination finds that standards spread when they wrap and bridge existing protocols rather than demand that everyone rewrite their systems to replace them Should coordination protocols wrap existing systems or replace them?. Agentic commerce shows the same pattern. ACP, AP2, UCP, Visa's TAP and Mastercard's Web Bot Auth don't really compete with each other. They form a ladder of layers covering application, governance, discovery and identity. Merchants stack them instead of picking one Do agentic commerce protocols compete or complement each other?. No single company can own a layered stack like that, which is exactly why it reshapes the market.

Protocols also need conditions they can't create by themselves. The UK competition regulator (the CMA) argues that data portability, meaning your ability to take your data with you, is basic market infrastructure. Without it, agents can't really act for consumers and switching stays blocked by incumbent lock-in Should data portability be infrastructure for agent-driven markets?. An open protocol on top of locked-up data still leaves the dominant actor in charge. The flip side is that platforms keep finding ways to profit even inside open-looking setups. One formal model shows a platform pricing "inspection tokens" (what a buyer pays to look at sellers) so that buyers over-search. Value moves from sellers to the platform without helping consumers Does platform pricing of inspection tokens encourage excess search?.

Protocols also lose some things that platform-controlled interfaces provide. Practitioners report that MCP caused unpredictable failures in production, such as agents picking the wrong tool or guessing parameters. Teams went back to direct function calls, and 85% of surveyed teams build custom agents instead of using frameworks Why do protocol-based tool integrations fail in production workflows?. Regulation runs into a similar problem. Busch argues that EU consumer law was built around screens, with mandatory buttons and disclosures. That law becomes impossible to enforce once agents transact with each other through protocol-layer APIs. His proposed fix is to write the obligations into the protocols themselves, for example a standard "withdrawal API" for cancelling purchases Can EU consumer law rules survive agent-to-agent API transactions?.

So what makes protocols market-shaping is that they move control from one owner into a shared layer that composes with others. That same move takes away the predictability that platforms provided and the screen-based levers that rules relied on. Whoever writes the protocol is now quietly writing the market's rules. The corpus is thinner on direct head-to-head comparisons. It has no empirical study measuring competition under an open protocol versus a dominant API in the same market, so the strongest claims here are still arguments, not measurements.