Just weeks after Meta poured billions into Scale AI, the American data-labeling startup is undergoing a major shakeup. The company announced it will lay off approximately 200 full-time employees — around 14% of its workforce — and let go of an additional 500 contractors globally. The decision follows a broader internal restructuring effort and growing concerns over recent data security lapses.

Confirming the layoffs, Scale AI spokesperson Joe Osborne said, “We are streamlining our data business.” The restructuring appears to be a strategic shift in response to rapid expansion that proved unsustainable.

In an internal email addressed to employees and reviewed by The Verge, Scale AI CEO Jason Droege explained the rationale behind the decision. “The reasons for these changes are straightforward: we ramped up our GenAI capacity too quickly over the past year,” he wrote. “While that felt like the right decision at the time, it’s clear this approach created inefficiencies and redundancies. We created too many layers, excessive bureaucracy, and unhelpful confusion about the team’s mission. Shifts in market demand also required us to re-examine our plans and refine our approach.”

Founded in the U.S., Scale AI plays a pivotal role in helping major tech companies like Google, OpenAI, Anthropic, and Meta by labeling and processing training data for their AI systems. Much of this work is carried out by contractors based outside the United States.

The announcement follows Meta’s substantial $14.3 billion investment in Scale AI, securing a 49% stake in the company. As part of the deal, Meta brought back former Scale AI CEO Alexandr Wang to lead its new superintelligence lab. This new division is reportedly recruiting senior talent from leading AI labs. Despite Meta’s deep involvement, Scale AI maintains that it continues to operate as an independent entity.

However, the timing of the layoffs has raised eyebrows, especially as the company continues to face backlash over recent security failings. Last month, a damning investigation revealed that at least 85 Google Docs containing sensitive and confidential client information were inadvertently left publicly accessible online.

Among the leaked documents were internal notes and training prompts for AI systems, including those for Google’s Bard (now Gemini), and confidential data from Meta and Elon Musk’s xAI. Shockingly, some documents also exposed personal information of contractors, including email addresses, payment disputes, and internal performance notes. One file titled “move all cheating taskers” openly listed names flagged for misconduct and was editable by anyone with access to the link.

While no major clients have publicly severed ties, reports suggest that several companies are reevaluating their relationships with Scale AI in light of the breach. According to a Business Insider report, some clients are growing increasingly cautious due to these lapses in data handling.

As Scale AI navigates this turbulent phase, it remains to be seen whether its restructuring and Meta-backed ambitions can rebuild client trust and realign its mission in the evolving AI landscape.