Runlayer and Rippling have ended their legal fight, but the dispute leaves behind a sharper lesson than the lawsuits themselves. On Wednesday night, both companies dropped their respective claims, with no settlement, no money changing hands, and not even lawyers’ fees exchanged, according to court documents seen by TechCrunch.
The product at the center of the fight was an MCP gateway, a piece of AI infrastructure that helps enterprises control how AI agents reach into business systems. Rippling marked the end of the dispute by releasing its own MCP gateway, a product that competes with Runlayer’s offering.
Why the dispute mattered
Runlayer is an early-stage startup that launched out of stealth in November, 2025. It has raised a total of $42 million from investors including Khosla Ventures’ Keith Rabois and Felicis.
The company is led by Andrew Berman, a third-time founder. His earlier companies include baby-monitor maker Nanit and Vowel, an AI video conferencing tool that sold to Zapier in 2024.
According to Runlayer’s lawsuit, Rippling tested Runlayer’s MCP gateway for more than a year. The two engineering teams worked closely during that period, but Rippling never became a customer. Instead, Berman received a text from a Rippling employee saying Rippling was building its own MCP gateway and planned to release it as a product. The employee described Rippling’s product as a clone of Runlayer’s.
Runlayer sued, arguing that Rippling had violated contractual agreements tied to the product tests. Rippling then countersued, alleging that Runlayer was violating some of its patents. Runlayer viewed that move as an attempt to push it into dropping its suit while increasing legal costs.
What an MCP gateway actually does
The dispute was not over a generic software feature. It centered on a layer that could become important as enterprises put AI agents to work inside existing software systems.
An MCP gateway securely manages an AI agent’s request for data from other business software. The basic idea is that an agent should not be handed direct access to every internal system it might need to query. Instead, the gateway sits between the agent and those systems.
For example, if a hiring professional asks for details on the top five candidates for a job, including their emails, that information has to come from the company’s recruitment system. The gateway handles that retrieval process.
That same gateway can also support other enterprise controls, including:
- Employee role-based access control, so managers and interns do not receive the same access.
- Observability, including logs and usage trails.
- A controlled path for AI agents to request business data without direct access to company software systems.
That explains why the category has drawn attention. An MCP gateway is not only about connecting AI agents to data. It is also about deciding who can access what, how requests are monitored, and how companies keep AI activity visible to IT teams.
The lawsuits ended without a deal
After spending the last three weeks in discovery, Runlayer dropped its suit. Rippling dropped its own suit as well. According to the TechCrunch report, there was no settlement and no payment by either side.
The short life of the lawsuits makes the outcome unusual but still useful for founders to study. The dispute did not end with a courtroom ruling, a licensing arrangement, or a public financial agreement. It ended with both sides walking away from the claims while Rippling moved ahead with its product release.
That means the broader question remains unresolved in a practical sense: how should startups protect themselves when large companies evaluate their products for long periods, work closely with their teams, and then decide to build something competitive?
The source article frames the episode as a cautionary tale for founders, especially in AI. Building new software has become easier, and a company that looks like a potential buyer can later become a direct rival.
Why AI product evaluations are getting riskier
Enterprises often put startups through technical evaluations before buying. Those evaluations can run for months and involve deep collaboration between engineering teams. In more stable software markets, that process can be frustrating but predictable.
AI infrastructure is moving faster. The needs and priorities of an enterprise can change substantially between the start of a product test and the end of it. A startup may enter the process expecting a customer relationship, while the enterprise may later decide that the same category is strategic enough to build internally.
That is the central lesson from the Runlayer and Rippling fight. The issue is not only whether contracts were strong enough, or whether lawsuits were the right response. It is that the commercial shape of an AI deal can change while the evaluation is still underway.
For founders, that creates several practical risks:
- A product trial can expose how a startup solves a problem before a customer commits.
- A long evaluation can give a larger company time to reassess whether to buy or build.
- A startup may spend months supporting a prospect that never signs.
- A product category can become crowded quickly as adjacent companies move in.
Rippling’s expanding AI infrastructure push
Rippling has historically been associated with payroll and benefits management. In the span of weeks, it has moved into the AI Gateway market with a tool that can route to different models while dashboarding token spend by employee. That product competes with the likes of Stripe, Ramp and Databricks.
With its MCP gateway, Rippling is also entering AI security. The product ties AI access to employee roles and competes with the likes of Runlayer, Docker and Amazon Bedrock.
Runlayer’s own pitch is broader than a single gateway function. It offers a bundle of agent security services tied to the gateway, ranging from agent creation to spotting shadow AI agents running inside an enterprise without IT knowing.
The dispute may be over legally, but the market conflict is not. Runlayer and Rippling are now part of a wider contest over how enterprises will connect AI agents to internal systems, monitor usage, control access, and manage security as AI tools spread through the workplace.
For AI startup founders, the practical message is blunt: product access, technical collaboration, and customer discovery all carry new weight when the customer may be capable of becoming a competitor before the deal is signed.