A federal court’s dismissal of xAI’s trade secret case against OpenAI is now headed to the Ninth Circuit, with Elon Musk’s company appealing a ruling that apparently found its claims over data center-related practices didn’t hold up, according to Datacenter Dynamics. The case matters beyond the Musk-vs-Altman rivalry: it’s one of a growing number of disputes testing whether the operational know-how behind frontier AI training — data center design, capacity planning, compute procurement — can be protected as a trade secret, or whether courts will treat it as the kind of general industry knowledge that moves freely with employees and competitors.
OpenAI’s move to recover its legal fees from xAI is the more consequential signal here. Fee-shifting in trade secret litigation typically requires a court finding that the underlying claim was brought in bad faith or was objectively baseless — a much higher bar than simply losing. If OpenAI succeeds, it would hand AI labs a template for punishing rivals who file trade secret suits as competitive pressure tactics rather than genuine IP protection, and it would raise the cost of bringing marginal claims across the sector.
For data center operators and AI infrastructure vendors watching from the sidelines, the underlying tension is familiar: as compute buildouts become the primary battleground between frontier labs, the engineering and sourcing details behind them are increasingly treated as competitive secrets worth litigating over, not just operational trivia. A Ninth Circuit ruling — whichever way it breaks — will shape how aggressively AI companies can invoke trade secret law to police departing engineers, partners, and rivals over data center know-how. Expect both sides’ briefing to focus heavily on what, if anything, actually qualifies as a protectable secret in an industry where scaling techniques diffuse quickly.
And OpenAI attempts to get its legal fees paid by Musk's company
