A federal judge has ruled that Palantir Technologies Inc. [NASDAQ: PLTR] must send its trade secret lawsuit against three former engineers to arbitration rather than proceeding in federal court, delivering a procedural setback to the company’s effort to stop former employees from using its proprietary technology to build a competing AI analytics company.
US District Judge Paul Oetken, sitting in New York, ordered arbitration on May 12, 2026, rejecting Palantir’s argument that seeking only injunctive relief entitled it to bypass the arbitration clauses in its employment agreements with the former staff.
Oetken wrote that the contractual exception Palantir relied upon applied solely to the enforcement of arbitration itself rather than to underlying employment-related disputes, a legal distinction that closed the avenue Palantir had pursued to keep the case before a federal judge.
The lawsuit names CEO Hirsh Jain, co-founder Radha Jain, and Joanna Cohen as the defendants, alleging they accessed Palantir’s source code and customer data during their employment and then breached non-disclosure and non-compete agreements when they departed to build Percepta AI, a competing AI analytics platform.
The defendants had pushed for arbitration in March, citing employment agreements that explicitly required such disputes to be resolved outside of federal court, a standard clause that has now been enforced against Palantir despite the company’s attempt to frame the relief it sought as sufficiently distinct to escape it.
Palantir had hoped that seeking an injunction to prevent the former engineers from using its proprietary information would be treated differently from a damages claim, but Oetken’s ruling made clear that the nature of the relief requested does not change the contractual obligation to arbitrate employment-related disputes.
The practical effect of the ruling is that the dispute will now proceed in private arbitration rather than federal court, which typically moves faster and with more limited discovery rights, and whose proceedings and outcomes are generally confidential, reducing the public record available about the specific technologies and customer data allegedly misappropriated.
The case adds to a broader pattern of AI talent disputes at frontier technology companies, where the rapid commercial value of proprietary systems and the high mobility of AI engineers frequently produce post-employment conflicts over intellectual property rights.
