On Wednesday night, Runlayer and Rippling dismissed their respective lawsuits against each other with no settlement, no monetary exchange, and no reimbursement of attorneys' fees, according to court documents reviewed by TechCrunch. Rippling immediately launched its MCP gateway product — the very technology that sparked the competing legal claims and that directly rivals Runlayer's service. The short-lived fight highlights a cautionary lesson for startup founders: in the AI era, when creating new software has become remarkably easy, your next rival could be anyone, even a potential customer.
The legal dispute unfolded after Rippling tested Runlayer's MCP gateway for over a year, with both companies' engineering teams collaborating closely throughout the evaluation period, yet Rippling never became a paying client. Runlayer is an early-stage company that emerged from stealth mode in November 2025 and has secured $42 million in total venture capital funding from investors including Keith Rabois at Khosla Ventures and Felicis. The startup is led by Andrew Berman, a third-time founder whose previous ventures include Nanit, a baby-monitor manufacturer, and Vowel, an AI video conferencing platform acquired by Zapier in 2024. Berman received a text message from a Rippling staff member stating that Rippling was developing its own MCP gateway and intended to launch it as a commercial product, with the employee characterizing Rippling's version as a duplicate of Runlayer's technology. Runlayer filed suit alleging Rippling breached contractual terms governing the product testing arrangement, while Rippling responded with a countersuit claiming Runlayer infringed on certain patents it holds. Runlayer withdrew its lawsuit after three weeks of discovery proceedings, and Rippling subsequently dropped its patent claim without collecting any financial compensation.
According to the report, an MCP gateway securely manages an enterprise's AI agent data requests from other software platforms — for example, when a recruiting manager asks for information on the top five job candidates including their email addresses, that information must be pulled from the company's hiring system. The gateway conducts the retrieval operation instead of allowing AI agents direct entry to the organization's software infrastructure, and it can add additional capabilities like role-based access controls that differentiate permissions for managers versus interns, observability features including logs and usage records, and related functions. Runlayer perceived Rippling's countersuit as a tactic to pressure it into abandoning its original claim while increasing its legal costs, according to the article.
The broader implication for founders, as the report explains, is that the AI sector is evolving so quickly that the extended technical evaluations enterprises traditionally require from startups need reconsideration. Between when an AI startup begins such an assessment and the conclusion months later, a company's requirements and preferences can shift dramatically. In just a matter of weeks, Rippling — whose core business has traditionally centered on payroll and benefits administration — entered the AI gateway market with a tool capable of routing requests to various models while providing dashboards showing token usage by individual employee. The product now competes with offerings from Stripe, Ramp, and Databricks. With its MCP gateway that links AI access permissions to employee roles, Rippling has also entered the AI security industry, positioning itself against competitors including Runlayer, Docker, and Amazon Bedrock. Runlayer's value proposition centers on a more comprehensive package of agent security services connected to its gateway, spanning from agent development to identifying shadow AI agents operating within an enterprise without the IT department's awareness. For enterprise buyers evaluating AI infrastructure, the speed at which established vendors can pivot into new categories may accelerate procurement timelines, while startups face heightened risk that prolonged evaluation cycles become competitive intelligence exercises rather than genuine sales processes.