Startup Litigation Vanishes Without Settlement, Exposing Rapid AI Landscape Shifts

Early Wednesday evening, Runlayer and Rippling abruptly ended a high-profile legal dispute that had captured the tech world's attention. Neither company paid the other, and per court filings reviewed by TechCrunch, legal costs went unrecovered. That same night, Rippling immediately unveiled its own MCP gateway—the very product central to the lawsuit.

The tangle between these two companies offers a stark reminder for founders navigating today's hypercompetitive AI landscape: the nature of enterprise relationships has fundamentally shifted.

How the dispute unfolded#

Runlayer, a security-focused startup founded by Andrew Berman (known for building baby-monitor company Nanit and video conferencing tool Vowel before its 2024 acquisition by Zapier), emerged from stealth last November with $42 million in backing from heavyweight investors including Khosla Ventures' Keith Rabois and Felicis.

The company's core offering is an MCP gateway—a secure middleware that mediates how AI agents request data from enterprise systems. Rather than giving agents direct access to sensitive databases, the gateway acts as a gatekeeper, managing role-based permissions, audit trails, and other governance layers. It's a critical piece of enterprise AI infrastructure.

For over a year, Rippling's engineering team tested Runlayer's technology closely. Then, without becoming a paying customer, a Rippling employee reportedly texted Berman to say the company would build its own competing product. Runlayer alleged the version Rippling developed mirrored its own, and sued for breach of contractual terms governing the evaluation period.

Rippling countersued, asserting patent violations. The move appeared calculated to pressure Runlayer into backing down and racking up defense costs. After three weeks of discovery, both sides withdrew their cases.

The real lesson: Enterprise velocity has changed#

The litigation came and went almost as quickly as it started, but the underlying dynamic it exposed is profound. Once, enterprises would spend months or years evaluating third-party solutions before committing. That evaluation phase offered startups breathing room.

No longer. Today's AI market moves so fast that by the time a company finishes testing a solution, its own engineering team may have already designed a replacement. The cost and speed of building new AI tools has collapsed, shifting incentives toward internal development over acquisition.

Rippling's pivot exemplifies this. Historically a payroll and HR platform, the company has now entered two new markets within weeks: AI model routing (where it competes with Stripe and Ramp) and AI security governance (directly challenging Runlayer, Docker, and Amazon Bedrock). Runlayer's own pitch spans a broader security suite covering agent creation, shadow AI detection, and access controls.