When AI infrastructure meets a "crisis of trust," the competition between big companies and startups is escalating.

The startup focusing on the Model Context Protocol (MCP) security gateway, Runlayer, has recently filed a lawsuit in court, accusing human resources software company Rippling of stealing its product ideas and trade secrets. This legal dispute reveals the intense competition and potential risks behind enterprise technology procurement in the AI era.

According to the complaint, the two parties had engaged in deep engineering collaboration and product trials for nearly a year. During this time, Runlayer shared its core product roadmap and underlying source code with Rippling, a potential customer. Both parties signed a non-disclosure agreement and a trial agreement containing intellectual property protection clauses. However, the testing was terminated as the two parties failed to reach an agreement on the purchase price.

Not long after, Runlayer's founder and CEO Andrew Berman received information from an internal source that Rippling was secretly developing a clone project almost identical to the startup's product. Based on this, Runlayer believes that the other party is suspected of infringing on trade secrets, breaching the contract, and engaging in unfair competition.

In response to the allegations, Rippling confirmed that it is launching its own MCP gateway but denied the accusation of abusing intellectual property rights, stating that its product is based on its own technological development.

With Anthropic releasing an open-source protocol, the MCP gateway has gradually become a key infrastructure connecting AI models with external data. However, as market competition intensifies, many large companies with strong engineering capabilities often choose to develop in-house, which puts many AI infrastructure startups relying on financing into a difficult dilemma.