Bloomberg reported in July that the U.S. data labeling startup Scale AI is undergoing significant adjustments, announcing layoffs of approximately 14% (about 200 people) and terminating partnerships with 500 global contractors. This star company, which once provided high-quality training data for many artificial intelligence labs, is experiencing a profound strategic transformation.
This adjustment comes after a major personnel change — last month, Meta hired Scale AI's former CEO Alexandr Wang at a valuation of $1.43 billion, sparking speculation about the company's future direction. Interim CEO Jason Droege admitted in an internal memo that the company's core data labeling business had expanded too quickly and was no longer suitable for the current market pace.
The memo stated that Scale AI will focus on strengthening its enterprise and government sales departments, concentrating on higher-value B-end services. This means the company will gradually move away from its once-proud data labeling "production line" model.
This trend is similar to recent strategic adjustments by several AI startups after "reverse acquisitions." For example, after Microsoft acquired Inflection AI, its technology and team were integrated into the large company's system, and its original product lines were gradually marginalized.
Notably, it has been reported that Meta's involvement caused some major clients of Scale AI to terminate their cooperation. As a data service provider, Scale AI faces not only changes in AI model training methods but also a loss of trust and neutrality in its partnerships.
Amid the ongoing surge of generative AI, the data labeling industry, which once provided fuel for large models, is now facing structural shocks. Scale AI's recent strategic retrenchment may also signal that an entire industry is entering a period of consolidation.