AI Gateway Grab Explodes in Runlayer Rippling Lawsuit

Runlayer's MCP gateway fight with Rippling is really a fight over who controls the access layer between AI agents and business software. The Runlayer Rippling lawsuit alleges that Rippling evaluated Runlayer’s product for nearly a year, saw confidential material under contract, then moved to launch its own competing MCP gateway, according to TechCrunch.

Runlayer’s claim is blunt: Rippling acted like a prospective customer, learned enough during a trial, failed to agree on price, then built what Runlayer says was effectively the same product. Rippling says that’s false, and that its own product was built with proprietary information.

The case lands at a sensitive moment for enterprise AI. Model Context Protocol, launched by Anthropic as an open source protocol in November 2024, gives AI models and agents a way to connect with outside data sources and services. MCP gateways sit in the middle, adding control, security, and management features. That middle layer is becoming valuable fast.

Runlayer's lawsuit turns MCP gateways into the next enterprise AI battleground

Runlayer, which offers a secure Model Context Protocol gateway, says Rippling got far more than a sales deck. In the complaint reviewed by TechCrunch, Runlayer describes “nearly a year of intensive engineering collaboration” during Rippling’s product trial.

That matters because enterprise software trials often require deep disclosure. Buyers want to test the product against real workflows, real constraints, and real technical requirements. Sellers want to close the deal. The danger is obvious: a technical buyer with enough engineering capacity may decide it can build instead of buy.

Runlayer says the parties signed a mutual non-disclosure agreement. Rippling also signed a product trial agreement that allegedly barred it from copying Runlayer’s intellectual property or making derivative works. Those clauses are standard in enterprise software trials, but this dispute shows why boilerplate suddenly carries real strategic weight in AI infrastructure.

Rippling’s position is equally direct. A spokesperson told TechCrunch:

“Runlayer’s panicked effort to avoid competition by fabricating claims is not an effective way to deal with its business failures.”

Rippling also confirmed to TechCrunch that it is launching its own MCP gateway. Its denial turns the case into a core question for the AI infrastructure market: where does legitimate internal development end, and trade secret misuse begin?

How Runlayer says Rippling moved from evaluation to imitation

The alleged sequence is simple. Rippling evaluated Runlayer’s MCP gateway as a prospective customer. During that process, Runlayer says it shared material including its product roadmap and actual source code. The two sides then failed to agree on price, and Runlayer ended the product trial.

Shortly afterward, Runlayer alleges that a “Rippling insider” texted founder and CEO Andrew Berman about an internal effort to copy the product.

“There’s been a project internally to build essentially a clone o[f] Runlayer … it’s almost a 1 to 1 copy of Runlayer.”

That text, if authenticated and placed in context, could become a central piece of the case. But it doesn’t settle the matter by itself. Trade secret cases usually turn on tighter evidence: what was disclosed, who accessed it, what contractual limits applied, whether the accused product reflects protected material, and whether the defendant can show independent development.

Runlayer’s legal claims include trade secret misappropriation, unfair competition, and breach of contract. The company has retained Sullivan & Cromwell, a major law firm. That doesn’t prove the case has merit, but it raises the optical stakes.

The filings that matter next will likely focus on:

• Contract scope: What exactly did the NDA and trial agreement forbid?
• Disclosure trail: Which documents, code, architecture details, and roadmap materials did Runlayer share?
• Access records: Which Rippling employees saw what, and when?
• Internal messages: Did Rippling discuss building a competing product using trial materials?
• Technical comparison: Does Rippling’s gateway resemble protected implementation details or just the same broad market idea?

That last distinction is crucial. Copying a product category is not the same as stealing trade secrets.

The numbers behind Runlayer's MCP gateway fight

The Runlayer Rippling lawsuit has enough hard numbers to show why both companies care.

These figures point to the real issue. MCP is open source, but enterprise-grade control around MCP is not free or generic. A gateway that manages how agents connect to business data and services can become infrastructure rather than a feature checkbox.

XOOMAR analysis: that is why the build-versus-buy decision is so tense here. For a large software company, buying or licensing a young vendor’s gateway could speed up launch. Building internally could protect product direction, customer data handling, and margins. For a startup, the same trial that proves value can expose enough detail to make the buyer less dependent on the vendor.

That tension resembles other cost-and-control decisions across tech, where internal capability and external vendors collide. XOOMAR has covered related pressure around automation and operating discipline in AI Shrinks Product Teams as Visa Layoffs Cut 2,600 and Cash Burn Clouds Tesla Q2 Earnings After $28.2B Haul. The specific facts differ, but the management question rhymes: what should a company buy, and what should it own?

Rippling, Runlayer, customers, and investors all want a different outcome

Runlayer’s likely view is that enterprise sales require trust. If a startup must reveal roadmap details, architecture, and code to win a customer, then confidentiality agreements are not paperwork. They are the sales model.

Rippling’s likely view is that evaluating vendors does not prevent it from building its own product. Companies routinely assess tools, learn what customers need, and then make internal roadmap decisions. That is lawful unless protected information was misused.

Customers sit in the middle. They want AI tools connected to useful business data, but they also want stable vendors, clear security controls, and products that will not be frozen or reshaped by litigation. A lawsuit around a core access layer can complicate procurement even when the technology works.

Investors in AI infrastructure should pay attention. Runlayer has raised $42 million, but capital does not remove the structural risk of selling to larger platforms. XOOMAR analysis: startups in this category may need tighter staged demos, cleaner audit trails, and stricter limits on source-code exposure during pilots. The more strategic the product, the more dangerous a casual trial becomes.

From open protocol to proprietary gateway, the ownership line is blurry

MCP creates a tricky legal and commercial boundary. The protocol itself is open source. The gateway products around it are where vendors add differentiated controls, security features, deployment patterns, and agent management.

That makes ownership claims harder to draw. A company cannot claim the broad idea of connecting AI agents to external systems through MCP. But it can claim specific protected implementation details, confidential architecture, source code, or trade secrets shared under contract.

This is the recurring platform problem in a new AI form. Startups often build the first useful layer around a fast-growing technical standard. Larger software companies then absorb similar functions into native products. Sometimes that is normal competition. Sometimes discovery shows something uglier.

The outside observer cannot know which version applies here yet. The complaint gives Runlayer’s account. Rippling denies misuse and says it is launching a superior product based only on its own proprietary information. The court process will decide whether the similarity, if any, reflects market convergence or misappropriation.

MCP gateway lawsuits could change how AI startups sell to big platforms

The practical lesson for founders is immediate. If your product sits between AI agents and enterprise data, assume your buyer may also be a future competitor.

That changes the sales motion:

• Demos: Show value without exposing implementation too early.
• Trials: Stage access, with technical gates tied to commercial progress.
• Contracts: Make anti-copying and derivative-work language specific, not decorative.
• Records: Track every document, code sample, architecture diagram, and meeting.
• Escalation: Treat pricing breakdowns as risk moments, not just negotiation failures.

For enterprise buyers, the opposite pressure appears. More restrictive trials can slow evaluations. Vendors may refuse broad technical access without stronger commitments. AI middleware procurement could become more legalistic, especially when the buyer has the talent to build a rival product.

XOOMAR’s position: novelty will not protect AI middleware startups for long. If an integration pattern becomes valuable, large platforms will move toward it. The defensible layer has to be specific: security depth, deployment expertise, trust, governance, cross-platform neutrality, and evidence that the product is more than a thin wrapper around an open protocol.

The Runlayer Rippling lawsuit previews harder fights over agent infrastructure

The Runlayer Rippling lawsuit is an early warning for the MCP gateway market. The core dispute is not whether Rippling can launch an MCP gateway. It can compete. The question is whether it used confidential Runlayer material to get there.

The strongest evidence for Runlayer would be specific: protected technical materials, employee communications tying those materials to Rippling’s build, code or architecture overlap, and a timeline that undercuts independent development. The strongest evidence for Rippling would be the reverse: clean internal records, independent design work, and proof that its product came from proprietary information rather than trial disclosures.

Watch the next filings, especially any motion over a preliminary injunction and any fight over expedited discovery. If the court lets Runlayer dig quickly into Rippling’s internal messages and technical records, the case becomes more dangerous for Rippling. If Runlayer cannot connect access to copying, the suit starts looking more like a failed vendor relationship turned courtroom fight.

Either way, the strategic stakes are clear. The company that owns the gateway between AI agents and business systems may own one of the most valuable control points in enterprise AI.

The Stakes

• The lawsuit highlights growing competition over the access layer between AI agents and business software.
• Enterprise software trials may face more scrutiny when buyers can build similar products in-house.
• MCP gateways are becoming strategically important as companies adopt AI agents connected to business systems.