Runlayer, a startup that sells a secure gateway for the Model Context Protocol, filed suit against HR software company Rippling on July 28, 2026, alleging that Rippling used a year-long enterprise product trial as cover to clone Runlayer's product. Runlayer has raised $42 million to date, including from Khosla Ventures and Felicis, and has retained Sullivan & Cromwell to press the case. The complaint accuses Rippling of trade secret misappropriation, unfair competition, and breach of contract.
According to the complaint, the two companies signed a mutual non-disclosure agreement and a product trial agreement containing a standard clause barring Rippling from copying Runlayer's intellectual property or building derivative works. Over what Runlayer describes as nearly a year of intensive engineering collaboration, the startup shared its product roadmap and source code with Rippling engineers. Talks broke down over price, and Runlayer ended the trial.
Shortly afterward, Runlayer CEO Andrew Berman received a text message from what the complaint calls a Rippling insider. The message, quoted in the filing, tipped him off to an internal Rippling effort that closely tracked what his team had spent the past year showing off.
Key facts
- 01Runlayer, which has raised $42 million from investors including Khosla Ventures and Felicis, sued Rippling for trade secret misappropriation on July 28, 2026.
- 02Runlayer alleges Rippling copied its MCP gateway after a product trial involving nearly a year of engineering collaboration.
- 03Rippling confirmed it is launching its own MCP gateway but denies using any of Runlayer's intellectual property.
- 04Runlayer retained Sullivan & Cromwell to press the case.
- 05Anthropic launched the Model Context Protocol as an open source standard in November 2024.
The insider's message is the factual spine of the suit. Rippling has confirmed to reporters that it is indeed building an MCP gateway, but denies any use of Runlayer's IP.
Rippling's public response is unusually pointed for a defendant at the complaint stage, framing the suit as a competitive tactic rather than a legal grievance. Runlayer has not yet responded to Rippling's characterization of its business.
A spokesperson said the following on the record:
MCP gateways sit on top of the Model Context Protocol, the open standard Anthropic released in November 2024 to let AI models and agents securely reach external data sources and tools. In under two years the protocol has become one of the load-bearing pieces of agent infrastructure, and a market for commercial gateways — adding access control, observability, secrets management, and agent policy on top — has grown up around it. Runlayer launched its product in the middle of last year into what was then a thinner field; it is now crowded, with hyperscalers, security vendors, and developer-tools companies all shipping their own versions.
That crowding is what makes the case interesting beyond the two named parties. Enterprise AI infrastructure sales cycles are long precisely because prospective customers demand deep technical evaluations — source code walkthroughs, integration work, roadmap sharing. For a startup, every one of those trials is a bet that the customer will buy rather than build. When the customer is itself a well-capitalized software company with a large engineering org, the build option is always live.
The legal question will turn on whether Rippling's gateway can be shown to derive from what Runlayer disclosed under NDA, or whether Rippling built independently using, as it claims, only its own proprietary information. Clean-room defenses are winnable in trade-secret cases, but they are expensive and document-intensive, and the alleged insider text — if authenticated — gives Runlayer a concrete piece of evidence to build discovery around. Sullivan & Cromwell's involvement signals Runlayer intends to litigate rather than settle quickly.
For the broader MCP gateway market, the suit is a warning shot at both ends of the deal. Startups pitching AI infrastructure to enterprise buyers with in-house engineering teams now have a live example of what a worst-case trial outcome looks like, and will likely tighten what they hand over during evaluations. Buyers, in turn, will find that trials come with more friction: fewer source-level walkthroughs, more redacted architecture diagrams, tighter derivative-works language. Both sides lose some velocity, and that is the real cost of a market where the customer can credibly become the competitor.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
