Meta denies there is any data quality issue with Scale AI. Surge and Mercor have both kept silent about the public split. In response to questions, a Scale AI representative pointed back to Meta’s original investment announcement, which projected a robust partnership.
After losing major contracts with OpenAI and Google, Scale AI laid off 200 staff from its data division, with its new chief executive citing sudden changes in market needs. The company has since pivoted, landing big contracts with the government and exploring new markets.
Meanwhile, within Meta, tension is rising. One current employee in the Meta Superintelligence Labs shared that several recruits from Scale AI are not actually working with TBD Labs, the group building the most ambitious AI systems.
Recent arrivals from outside Meta, especially former OpenAI and Scale AI talent, have struggled with Meta’s size and complex corporate culture. Multiple well-known faces from Meta’s earlier AI team have left after finding their influence reduced.
Rishabh Agarwal, a respected MSL engineer, announced his exit on X this week. “The pitch from Mark and @alexandr_wang to build in the Superintelligence team was incredibly compelling,” Agarwal wrote, “But I ultimately choose to follow Mark’s own advice: ‘In a world that’s changing so fast, the biggest risk you can take is not taking any risk.’”
Amid the turbulence, Meta is plowing ahead with huge data infrastructure projects, including the construction of a fifty billion dollar data center in Louisiana called Hyperion. Work on Meta’s next generation of AI models is already underway, with an ambitious target launch before the year ends.
