Two Tech Companies Sued Each Other Over an AI Product, Then Both Walked Away Empty-Handed
Runlayer and Rippling dropped their respective lawsuits against each other on the night of August 19, 2026. No settlement. No money. Not even lawyers' fees changed hands, according to court documents reviewed by TechCrunch. Reuters confirmed the development the following day, with spokespeople for both companies acknowledging the end of the legal dispute.
The fight was about something called an MCP gateway. As companies start using AI agents (software programs that can carry out tasks on their own), those agents need a safe way to connect to a company's other software. An MCP gateway is a go-between. Instead of letting an AI agent connect directly to, say, a company's payroll system, the gateway sits in the middle, checking permissions and logging what the agent does. Runlayer built one. Rippling, a company known for payroll and benefits management, then built one too and released it.
Runlayer launched publicly in November 2025 and has raised $42 million from investors including Khosla Ventures' Keith Rabois and Felicis. The company is led by Andrew Berman, a third-time founder whose previous ventures were baby-monitor maker Nanit and AI video conferencing tool Vowel, the latter sold to Zapier in 2024. For a founder at that stage, taking on a well-funded adversary in court is a calculated bet, and one that, in this instance, produced no financial recovery.
The dispute goes back to a long evaluation period. According to Runlayer's lawsuit, Rippling tested Runlayer's MCP gateway for more than a year, with the two engineering teams working closely together. But Rippling never signed on as a customer. Runlayer says that during this time Rippling got a good look at its product and how it was built, and that a Rippling employee later texted Berman to say Rippling was building its own MCP gateway, planned to release it, and described it as a clone of Runlayer's. Rippling denied the trade-secret theft allegations.
Runlayer sued first, claiming Rippling violated contractual agreements covering the product tests. The original suit alleged trade secret theft, according to the New York Post. Rippling fired back on August 10, 2026, filing a patent-infringement lawsuit in Delaware federal court accusing Runlayer of infringing three of Rippling's patents. Runlayer saw the countersuit as a tactic to drive up legal costs and pressure the smaller company into dropping its claims.
That pressure, if it was the intent, appears to have worked. Runlayer had spent the last three weeks in discovery (the pre-trial phase where both sides exchange evidence) before dropping its suit. Rippling then dropped its own suit without collecting anything. Both companies walked away with nothing: no admissions, no payments, no licensing arrangements.
Rippling has since released its MCP gateway as a commercial product, competing directly with Runlayer's offering. The Rippling gateway routes to different AI models and tracks how much each employee spends on AI usage, tying that to the same workforce data it already uses for payroll and HR. For Rippling, the gateway extends its existing employee-data advantage into AI governance. For Runlayer, it means competing head-to-head with a company that had over a year of hands-on exposure to its product.
This pattern will be familiar to anyone who has watched the tech industry for long. A large company evaluates a smaller company's technology, then builds its own version. The smaller company sues. Legal bills pile up. The smaller company, running low on cash, drops the suit. The larger company releases its product anyway. The cycle is not new, but the MCP gateway category is, and the speed of this dispute — roughly nine months from Runlayer's public launch to the lawsuits being dropped — is faster than earlier enterprise software disputes.
The structural imbalance is the real story here. A company like Rippling can absorb a patent lawsuit as a cost of doing business. A startup with $42 million in funding cannot keep paying legal fees indefinitely without it eating into the money meant for building and selling its product. The countersuit — patent claims filed in Delaware federal court — is a well-known pressure point, and Runlayer's claim that it was designed to rack up costs is plausible whether or not the patent arguments have merit. Neither company took the claims to trial, so the patent questions remain untested.
What this means for the market is straightforward. There are now at least two competing MCP gateway products: one from a payroll-and-HR company using its existing workforce data, and one from an independent startup. The legal fight did not slow either product down. It did, however, cost Runlayer time and money it would have rather spent elsewhere, and it sets a precedent that other large companies evaluating similar startup technology will notice.
In this author's view, the episode is worth flagging for what it reveals about the current pace of AI development. New product categories are forming so quickly that the usual rules and norms around evaluation, partnership, and intellectual property have not caught up. The specific technology here is narrow, but the underlying story — a small company builds something useful, a larger company takes a close look, then builds its own version — is one we have seen before. What has changed is the speed, and that speed leaves startups with even less room to react when a bigger competitor moves.
