Now Rippling is counter-suing tiny startup Runlayer

Rippling just flipped the table. The $13 billion HR unicorn filed a counter-suit Monday accusing Runlayer — a startup that has raised $42 million total — of infringing three Rippling patents. This comes weeks after Runlayer sued Rippling for breach of contract and stealing product ideas. The symmetry is perfect. The asymmetry is the story.

Runlayer builds an MCP gateway. MCP, the Model Context Protocol, is an open standard. It lets AI agents talk to data and software systems. Runlayer bundles cybersecurity features on top: threat detection, anomaly monitoring. Founder Andrew Berman has done this before. Nanit made baby monitors. Vowel made AI video conferencing tools before selling to Zapier. He knows how to package infrastructure.

Rippling became an early trial customer. Nearly a year of testing. No price agreed. No contract signed. Then Rippling built its own MCP server and plans to sell it. This is the Rippling playbook: use a tool internally, then productize it. The AI Spend Console followed the same path.

Now zoom in on the allegations. Runlayer claims a Rippling employee reached out to Berman directly to warn him: your customer is building a copy of your product. That employee has since "revised that view," according to Rippling. The revision is the most honest sentence in this whole mess. An engineer saw what was happening. They said something. Then the lawyers got involved.

Rippling's counter-narrative: it informed Runlayer of the allegedly infringed patents soon after Runlayer filed suit. The timing suggests leverage. Runlayer calls it a "desperate, retaliatory ploy." Rippling calls Runlayer's original suit "manufactured claims to distract from business failures." Both statements were written by PR professionals paid to sound righteous. Neither reveals what the patents actually cover.

Here is the uncomfortable question: what does an MCP gateway patent even look like? The protocol is open. The spec is public. Anyone can implement it. Patents on top of open standards usually cover specific optimizations — connection pooling, authentication flows, routing logic. Valid? Maybe. Broad? Likely. Weaponizable? Absolutely.

Rippling has the war chest. Runlayer has the runway. That dynamic shapes every move. A lawsuit costs Runlayer existential money. For Rippling, it's a line item. The counter-suit raises the cost of Runlayer's original action. It signals: keep fighting and the meter runs faster. This is not about justice. This is about leverage.

The industry watches this pattern repeat. Big company trials small company's product. Big company learns the architecture. Big company builds replacement. Small company sues. Big company counter-sues with patent portfolio. The legal fees become the acquisition price. Or the small company folds. Either way, the big company wins.

Runlayer's product launched a year ago. That is not much moat. Cybersecurity features on an open gateway are differentiable but not defensible. Threat detection on MCP traffic is a feature, not a barrier. Rippling's internal version benefits from the year of watching Runlayer operate in production. That is the advantage no patent covers: observed reality.

Berman's statement claims a "standout AI product that has nothing to do with these patents." He may be right. But "nothing to do with" is a high bar in patent litigation. Claims construe broadly. Infringement finds a way. The courts will spend years parsing claim language while Runlayer burns cash.

Rippling's spokesperson matches the theater: "It takes a certain boldness to accuse a competitor of violating intellectual property laws while infringing on that competitor's inventions." The word "hypocrisy" appears. The phrase "business failures" appears. This is not legal argument. This is narrative warfare.

The employee who warned Berman is the only human in this story. Everyone else is a corporate avatar. That engineer saw the cloning in real time. They tried to stop it. The machine corrected them. That moment — the revision — captures the entire dynamic. Individuals see theft. Organizations see strategy.

Settlement is the only rational exit. Rippling gets a license or a buyout at distressed price. Runlayer survives. The lawyers get paid. The patents never get tested. The open standard keeps moving. The next startup watches and learns: don't trial with giants without a signed contract and a defined exit clause.

But the damage compounds. Startups will stop trusting enterprise trials. Enterprises will stop getting early access to innovation. The trial model — the engine of B2B software adoption — fractures. Everyone loses except the incumbents who can afford to build in-house.

The courts will unwind who did what to whom. Or they won't. The docket fills. The bills mount. Runlayer's $42 million evaporates in legal fees. Rippling's MCP server launches regardless. The product ships. The lawsuit becomes a footnote in the release notes.

This is the AI-powered age of fast product building. The building is fast. The copying is faster. The suing is fastest. The customer-startup relationship didn't devolve. It was always this fragile. We just pretended otherwise.