Patent Counterclaim Reveals the Hidden Cost of AI Collaboration

When the Trial Never Ends

Rippling filed a patent infringement complaint this week targeting Runlayer, a small MCP gateway provider, alleging violations across three of its patents. The move comes roughly one month after Runlayer initiated its own litigation, claiming breach of contract and misappropriation of proprietary concepts following an 11-month product evaluation that never converted to a paid relationship.

The collision between a well-funded HR platform and a 42-million-dollar startup serves as a cautionary data point for founders navigating extended enterprise pilots in the AI infrastructure era. Runlayer, founded by serial entrepreneur Andrew Berman, launched its MCP gateway product approximately one year ago, bundling connectivity features with integrated threat detection and cybersecurity controls. Model Context Protocol has emerged as a critical plumbing layer, enabling AI agents to interface with external data sources and software systems required for autonomous operation.

Rippling became an early trial participant, testing Runlayer's offering for nearly a year. According to filings, the two parties never reached agreement on commercial terms. Instead, Rippling developed its own MCP server internally and plans to commercialize the technology, continuing a pattern the company has established with other in-house tools such as its AI Spend Console released earlier this year.

The Whistleblower Detail

Among the more striking claims in Runlayer's original complaint is an allegation that a Rippling employee contacted Berman directly to warn him that the company was building what the employee characterized as a duplicate of Runlayer's product. Rippling has since stated that the individual in question has walked back that characterization, though the episode underscores the tension that can develop inside organizations when competitive boundaries become ambiguous.

Rippling disclosed the three patents it believes Runlayer infringes shortly after Runlayer filed its initial lawsuit. The timing suggests the counterclaim may function as much as a negotiating tool as a legal remedy, a dynamic both sides appear to recognize. Berman described the patent suit as a retaliatory distraction, emphasizing that Runlayer's product stands independent of the disputed intellectual property and that the company will continue to serve what he characterizes as a rapidly expanding customer base.

Rippling, for its part, framed the counterclaim as a response to what it views as manufactured allegations designed to obscure Runlayer's own business challenges. The company pointed to what it called hypocrisy in Runlayer's position, accusing a competitor of IP violations while simultaneously infringing on that competitor's inventions.

The Build-or-Buy Calculation Shifts

The dispute offers a window into how generative AI and agent frameworks are reshaping the economics of enterprise software procurement. Large organizations now possess tooling and talent that can accelerate internal development cycles, reducing the friction historically associated with building proprietary alternatives to third-party solutions. Extended evaluation periods, once viewed primarily as sales cycles, increasingly double as reconnaissance missions, with vendors exposing product architecture, API design, and feature roadmaps in the hope of closing a contract.

For startups, the risk calculus has changed. Offering a trial to a well-resourced enterprise customer can yield valuable feedback and a marquee logo, but it also grants that customer months of visibility into implementation details, performance characteristics, and competitive positioning. When the trial concludes without a signed agreement, the startup may find itself facing a newly launched competitor built by the same organization that requested the proof of concept.

Rippling's model of converting internal tooling into standalone products amplifies this dynamic. The company builds systems to solve its own operational needs, then packages those systems for external sale. In this case, an MCP server developed to support Rippling's own AI initiatives will soon compete directly with Runlayer in the market. Whether that development constitutes independent invention or impermissible appropriation will ultimately be decided by the courts, unless the parties reach a settlement first.

What the Case Signals for Founders

At DailyTechWire, we have tracked a growing number of disputes rooted in the tension between collaborative product development and competitive intelligence gathering. The Runlayer-Rippling conflict is notable not for its novelty but for how clearly it illustrates the structural incentives at play. Enterprises benefit from extended trials that de-risk purchasing decisions and provide optionality to build in-house if terms prove unfavorable. Startups, especially those operating in infrastructure categories adjacent to their customers' core competencies, face the challenge of demonstrating value without revealing the implementation details that make the product defensible.

Berman's background as a third-time founder adds another layer to the narrative. His previous ventures include Nanit, a baby-monitor hardware company, and Vowel, an AI-enhanced video conferencing platform acquired by Zapier in 2024. The experience navigating those exits likely informed Runlayer's decision to litigate rather than negotiate quietly, signaling to the broader market that extended trials carry legal risk for both parties.

The open-source nature of the Model Context Protocol further complicates the IP landscape. MCP itself is a standard, not proprietary technology, which means differentiation comes from implementation details, security features, and operational tooling built around the core protocol. Runlayer's bundling of threat detection with its gateway represents one such differentiation strategy. Whether Rippling's implementation crosses the line from inspiration to infringement will hinge on technical specifics that discovery and expert testimony will need to unpack.

The Settlement Pressure

Both companies have issued statements heavy on combative rhetoric, but the structure of the litigation suggests settlement remains the most likely outcome. Patent disputes in the software infrastructure space are expensive, time-consuming, and uncertain. For Runlayer, continued legal costs could constrain the runway provided by its 42-million-dollar raise. For Rippling, a protracted case risks discovery processes that could expose internal communications and decision-making around the MCP server's development, documentation that might prove awkward even if ultimately exculpatory.

The broader ecosystem is watching. Founders evaluating enterprise pilots now have a high-profile case study illustrating the downside risk of deep product engagement without contractual safeguards. Enterprises, meanwhile, face growing scrutiny over how they manage the boundary between legitimate evaluation and competitive intelligence. The AI agent category, still in its early formation, lacks the established norms and contractual templates that govern more mature software markets, leaving both sides to navigate ambiguous terrain.

As MCP adoption accelerates and more enterprises deploy agent-based workflows, the demand for secure, reliable gateway infrastructure will grow. Runlayer and Rippling are far from the only players in the space, but their legal confrontation may shape how other participants structure customer relationships, pilot agreements, and IP protections. For now, the case moves to discovery, where the technical and documentary evidence will determine whether Rippling independently invented its MCP server or whether Runlayer's allegations of misappropriation hold weight. Either way, the episode stands as a reminder that in the era of rapid AI-driven development, the line between collaboration and competition has never been thinner.