Scale AI slashes 14% of staff after massive Meta investment

What happens when a tech giant hands your startup a $14.3 billion lifeline and simultaneously poaches your visionary CEO?

At Scale AI, the answer is cutting 14% of the team just one month later.

The data-labeling heavyweight confirmed it is laying off 200 full-time staff, roughly 14% of its global workforce, while parting ways with about 500 of its thousands of global contractors. Full-time workers will receive severance.

The hangover from a mega-deal

The shakeup comes four weeks after Meta closed a massive $14.3 billion deal for a 49% stake in Scale. As part of that agreement, Scale co-founder and CEO Alexandr Wang left his own company to run Meta’s new superintelligence group.

In a memo seen by Bloomberg, interim CEO Jason Droege told staff that the layoffs were driven by the data-labeling arm hiring too many people too quickly over the past year. That rapid headcount growth led to excessive bureaucracy, too many management layers, and confusion over team priorities, alongside broader shifts in market demand.

But the elephant in the room is client trust.

Following Meta's quasi-acquisition, marquee clients including OpenAI and Google began quietly winding down their work with Scale. Neither competitor was thrilled about sending proprietary training workflows through a shop now tightly tied to Mark Zuckerberg. Rivals like Turing, Invisible Technologies, Labelbox, and Uber have already reported a surge in inbound interest.

Scale is not short on cash. The startup generated around $870 million in revenue in 2024 and projected $2 billion for 2025. Droege noted the company will tap its war chest to refocus the labeling unit on higher-value domains like coding, language, and audio.

Scale also plans to hire hundreds of new staff in the back half of 2025 for enterprise applications and government contracts, including work with the U.S. Department of Defense. Droege previously said those enterprise and public-sector operations are already pulling in nine figures in revenue.

Why it matters

Scale built an empire by acting as the neutral Switzerland of AI training data. The moment Meta bought nearly half the equity and took the CEO, that neutrality vanished. OpenAI and Google walking away was entirely predictable, proving that in the frontier AI race, conflict of interest trumps almost any vendor relationship.

This also spotlights the human cost of Big Tech's clever new regulatory playbook. Regulators frown on outright acquisitions, so giants like Meta and Google are structuring deals to license tech, take non-controlling stakes, and poach founding teams, as Google recently did with Windsurf. The founders get massive paydays, while the teams left behind face identity crises, client flight, and restructuring memos.

We think Scale’s pivot toward specialized coding data and defense contracts is practically mandatory now. Simple data labeling is increasingly commoditized or replaced by synthetic pipelines. To hit its ambitious $2 billion targets without its founding architect or its biggest frontier lab clients, Scale has to morph from a consumer tech vendor into an enterprise defense powerhouse.