The fight in Versata Software, LLC v. Ford Motor Company was never about whether Ford misappropriated trade secrets — the jury said it did, and the Federal Circuit left that finding alone. The fight was about the dollars, and on May 22, 2026 the appeals court reopened them. The panel vacated the district court's trade-secret damages judgment and remanded for a new trial, with explicit instructions to consider two damages models the trial court had thrown out. It also reversed the district court's near-total reduction of the contract award and reinstated the jury's $82,260,000 verdict. The commercial stakes are large; the doctrinal point is about what a trade-secret plaintiff is allowed to ask a jury to award.
The dispute traces to software Versata (formerly Trilogy) built for Ford to manage vehicle configuration — the Automotive Configuration Manager and an associated cost module. Versata licensed the software to Ford under a master agreement, and the relationship soured when Ford, in 2014, terminated and rolled out its own configuration software, PDO, developed while it still licensed Versata's product. Litigation followed, with Versata asserting misappropriation of three interdependent "combination" trade secrets under the federal DTSA and Michigan's UTSA, plus breach of contract. The jury found Ford liable on both, awarding $22,386,000 in trade-secret damages and $82,260,000 for breach.
"it measures Ford’s enrichment rather than Ford’s unjust enrichment."— Federal Circuit, source
That single sentence — the district court's rationale for excluding Versata's damages expert before trial — is what the Federal Circuit reexamined. The trial court had barred the expert's unjust-enrichment theory under Daubert, reasoning that the model captured Ford's enrichment rather than its unjust enrichment, would hand Versata "far more than the fair price" Ford should have paid, and would confer "a huge and undeserved windfall." It then confined Versata to a reasonable-royalty model tied to the parties' licensing history, admitting one such model and excluding two others that incorporated the value to Ford of using the secrets. The jury's $22 million trade-secret award was built on the licensing-history model alone — and even that the district court later reduced to zero on JMOL, finding the jury had no reliable way to determine how long it would have taken Ford to develop the three misappropriated secrets.
Unjust enrichment is not capped at a license fee
The damages theory is the heart of this case. The Federal Circuit's instruction on remand — to consider the two previously rejected models — signals that a trade-secret plaintiff's recovery is not necessarily limited to what the parties' prior licensing history would have produced. Unjust enrichment in trade-secret law is meant to capture the benefit the defendant wrongfully obtained, which can exceed a negotiated license rate; treating the two as interchangeable risks under-compensating the very harm the cause of action targets. The district court's concern about a "windfall" is a real one, and apportionment remains essential, but the appellate court's disposition makes clear that the answer is to test those models on a proper record, not to exclude unjust-enrichment theories at the threshold and funnel the plaintiff into a license-history royalty.
That said, the opinion is not a blank check. The Georgia-Pacific-style reasonable-royalty models Versata ultimately submitted were framed as a hypothetical 2011 negotiation, and apportionment and reliable data inputs were live issues throughout. The remand puts those questions back before the district court with the excluded models in play, which is where the contested expert work — enrichment versus unjust enrichment, the value to Ford of the misappropriated functionality, and how to apportion it across the three combination secrets — will have to be done properly.
The contract award comes back in full
On the breach-of-contract side, the court was more decisive. The jury had awarded $82,260,000 for Ford's breach of the master agreement, and the district court reduced it to $3, reasoning the jury had no way to calculate Versata's contract damages with reasonable certainty. The Federal Circuit reversed that reduction and reinstated the jury's award in full. The panel also affirmed the district court's denial of Ford's JMOL motion on trade-secret liability, so Ford's cross-appeal challenging the misappropriation finding failed.
It is worth being precise about the procedural posture, because it explains why two damages tracks ended up in such different places. The trade-secret award was shaped twice over: first at the Daubert stage, where the court excluded the expert's unjust-enrichment model and confined Versata to a licensing-history royalty, and again on JMOL, where the court zeroed out even the $22 million the jury had awarded on that single admitted model. The Federal Circuit's vacatur and remand reopen the front end of that chain — the exclusion — which is the more consequential ruling, because it determines which theories Versata may present at all on retrial. The contract award traveled a different road: the jury's $82,260,000 figure was reduced to a nominal $3 on a reasonable-certainty rationale, and the appellate court simply restored the jury's number. The contrast underscores a practical point about trade-secret cases: damages theory, not liability, was the real fight here, and the theory a plaintiff is permitted to put before the jury can swing the recovery by tens of millions of dollars.
For licensors of enterprise software, the case is a useful marker on remedies strategy. When a customer terminates and stands up a competing internal product built on knowledge gained under license, the plaintiff's damages case should preserve both tracks — a contract measure and a trade-secret unjust-enrichment measure — rather than assume the two collapse into a single licensing-history royalty. The reinstated $82 million contract verdict shows the contract track can stand on its own, and the remand shows the unjust-enrichment track deserves a real hearing. What the case does not do is decide the ultimate number; that will be litigated again, this time with the full menu of damages models the trial court had taken off the table.