Scale AI sues rival Mercor over data-labeling trade secrets in high-stakes battle

The dispute at a glance

A high-stakes legal clash has erupted in the AI data-labeling arena. Scale AI has filed a lawsuit in a California court against its rival Mercor and a former Scale executive, Eugene Ling, alleging a brazen scheme to steal confidential documents and lure a major client away from Scale. The complaint centers on Ling allegedly downloading more than 100 sensitive documents related to customer strategies before joining Mercor and actively pitching Mercor’s services to a significant client during the transition period.

This isn’t just a breach-of-contract squabble. Scale claims a web of trade-secret misappropriation and corporate espionage, arguing that Ling transferred Scale’s proprietary files to a personal Google Drive weeks before his departure. The documents reportedly touched on strategies for a large, unnamed client, referred to in court filings as “Customer A.” Scale alleges that Ling was soliciting this client on Mercor’s behalf before his official exit, a move the lawsuit describes as a serious breach of duty.

Scale is seeking damages, legal costs, and a court order to prevent Mercor from using the allegedly stolen information and to bar Ling from working with Customer A. The stakes are unusually high because the client contract in question could be worth millions, underscoring why the case has drawn so much attention in a market where data and talent are becoming the most valuable currencies.

We won’t allow anyone to take unlawful shortcuts at the expense of our business, Scale AI asserted in response to the allegations. The company framed the lawsuit as a defensive measure in an environment where competitive lines are increasingly blurred and where a single customer move can tilt the landscape.

The parties’ positions

Scale AI’s claims

• The 28-page complaint paints a picture of Ling operating in near real-time with Mercor while still employed at Scale, allegedly moving restricted files to a personal cloud storage and actively courting a major client for Mercor.
• It highlights the alleged timing: the weeks leading up to Ling’s departure, with documents tied to Customer A, and a plan to switch that major account to Mercor’s fold.
• Scale argues that a pre-emptive departure accompanied by the misappropriation of strategy documents signals a calculated effort to siphon a valuable client and undermine Scale’s competitive edge.

Mercor’s response

• Mercor co-founder Surya Midha publicly pushed back, stating that the company hired Ling and other former Scale employees but does not seek to use Scale’s confidential information.
• Midha contends Ling had “old documents in a personal Google Drive” and says Mercor never accessed them. He also notes that Mercor contacted Scale about a resolution six days before the lawsuit was filed, offering to destroy the files or work out a settlement, a proposal Scale reportedly rejected.
• Mercor positions itself as a disruptor in a market long dominated by Scale AI, focusing on matching elite human expertise with AI labs’ needs rather than chasing a bundle of confidential assets.

Market context and implications

The data-labeling market is a crucial but competitive layer in AI development. As models grow more capable, the industry relies on sophisticated labeling work to translate raw data into trainable signals. Mercor’s emphasis on hiring PhDs and specialized professionals signals a niche strategy designed to compete with Scale’s broader services. Scale’s recent corporate shake-ups—including a notable $14.3 billion investment from Meta that gave it a 49% stake in the company—add another layer of sensitivity. Some clients reportedly expressed concerns about conflicts of interest in the wake of the deal, intensifying the pressure to safeguard customer relationships and proprietary know-how.

These dynamics aren’t happening in a vacuum. The Scale-Mercor dispute mirrors other high-profile cases involving trade secrets and employee moves within AI and tech, such as Elon Musk’s xAI taking legal action over allegedly stolen information. In an industry where the most valuable assets are intangible—think proprietary data, training strategies, and trusted client relationships—the outcome of this lawsuit could set meaningful precedents for how trade secrets are protected and how employee transitions between rivals are managed.

Looking ahead

While the court battle unfolds, observers will watch not just who wins on the merits but how the case reshapes norms in the AI data-labeling ecosystem. If Scale succeeds, it could deter aggressive poaching and encourage more robust protections around customer data and internal playbooks. If Mercor maintains that it did not access confidential information and that Ling’s role did not involve misuse of Scale’s assets, the industry could see a reinforcement of more collaborative means to resolve disputes without eroding trust or stifling innovation.

The broader takeaway is a reminder that in the current AI gold rush, the most valuable assets aren’t just code or models—they’re the carefully guarded combinations of data, strategy, and client trust that power real-world deployments.

Next steps

As the case proceeds, expect a string of filings, depositions, and possibly settlements that could ripple through the sector. Legal outcomes here can influence how companies design employee transitions, what constitutes protectable trade secrets, and how agreements with major clients are structured in a landscape where competition is fierce and the stakes are high.