Runlayer’s lawsuit against Rippling has turned a dispute over one enterprise software trial into a wider signal for the AI infrastructure market. The case centers on a secure Model Context Protocol gateway, a product category built around helping AI models and agents connect to outside data and tools.
According to the complaint seen by TechCrunch, Runlayer says Rippling received deep access during a product evaluation and later began preparing its own MCP gateway. Rippling confirms it is launching that product, but rejects Runlayer’s allegation that it misused the startup’s intellectual property.
A trial that became a legal fight
Runlayer offers a secure Model Context Protocol gateway. MCP is described in the source article as a standard that lets AI models and agents securely pull in outside data and tools.
The lawsuit says Rippling entered a broad product trial as a potential customer. During that process, Runlayer says it shared sensitive material, including its product roadmap and its actual source code.
The companies signed a mutual non-disclosure agreement. Rippling also signed a product trial agreement that included a clause barring it from copying Runlayer’s intellectual property or creating derivative works. The source describes that kind of language as standard boilerplate in enterprise software trials.
Runlayer says the evaluation involved “nearly a year of intensive engineering collaboration.” The two sides ultimately did not agree on a price, and Runlayer ended the product trial.
What Runlayer alleges
After the trial ended, Runlayer claims it received a warning from inside Rippling. The complaint says a “Rippling insider” texted Runlayer founder and CEO Andrew Berman about “a project internally to build essentially a clone o[f] Runlayer … it’s almost a 1 to 1 copy of Runlayer.”
Runlayer argues that Rippling’s product must have relied on Runlayer’s intellectual property. In the suit, it brings claims tied to trade secret misappropriation, unfair competition and breach of contract.
The startup has retained Sullivan & Cromwell. The source article notes that this does not determine whether Runlayer will win, but says a prominent law firm can add optical credibility to a lawsuit in much the same way a prominent venture firm can add credibility to a startup.
Rippling’s response
Rippling told TechCrunch that it is launching its own MCP gateway. At the same time, a company spokesperson denied Runlayer’s claims about misuse of intellectual property.
“Runlayer’s panicked effort to avoid compet ition by fabricating cla ims is not an effective way to deal with its business failures. Rippling is launching a superior product for connecting AI tools to business data using only our proprietary information – we have every reason to win in this market,”
That response puts the dispute in direct terms. Runlayer frames the case as a misuse of confidential access granted during a product trial. Rippling frames its product as an independent launch built from its own proprietary information.
Why MCP gateways are under pressure
The backdrop is a fast-moving MCP gateway market. Anthropic launched MCP as an open-source protocol in November 2024. Since then, the source article says it has become one of the basic building blocks of AI interoperability.
MCP gateway products add control, security and other features, particularly for managing agents. Those capabilities matter because AI models and agents increasingly need structured ways to reach external data sources and services.
Runlayer launched its product in the middle of last year and has raised a total of $42 million, including from Khosla Ventures and Felicis. But the source also describes the MCP gateway category as increasingly crowded.
That competitive pressure matters for startups selling AI infrastructure. A product can be useful enough to trigger a long technical evaluation, while also sitting in a category that larger or technically sophisticated customers may decide to build themselves.
The enterprise AI sales problem
The case also highlights a practical tension in enterprise AI sales. Complex infrastructure often requires deep, hands-on trials before a buyer is ready to commit. For a startup, that can mean giving a prospective customer detailed access to the product, roadmap and implementation.
For the customer, the same trial can become a learning process. The company may discover what it needs, what it does not need and whether its own engineers can build an internal version instead of buying from a vendor.
That creates a difficult balance for both sides:
- Startups need access. Without serious technical collaboration, enterprise customers may not trust the product enough to buy it.
- Customers need evaluation. They often have to test security, integration and workflow fit before committing.
- Competitive risk remains. A company with enough engineering capacity may decide to build rather than buy after seeing what the product does.
Runlayer’s lawsuit does not settle that tension. It does, however, make the risk visible. In a market where MCP gateways are becoming more competitive, the line between legitimate evaluation and disputed copying can become a central business risk.