OpenAI shifts focus: building advanced AI models after phasing out Scale AI

OpenAI is officially phasing out its partnership with Scale AI, a data-labeling startup, marking a strategic shift in its AI development approach.

The decision comes shortly after Meta Platforms Inc. invested $14.3 billion in Scale AI and recruited its CEO, Alexandr Wang, to lead a new "superintelligence" unit aimed at creating more powerful AI software.

"Scale accounted for a small fraction of OpenAI’s overall data needs," an OpenAI spokesperson said, confirming the move.

This transition reflects OpenAI's focus on building more advanced AI models capable of mimicking human reasoning processes and developing agent-like models that can autonomously perform tasks with minimal user input.

Meta’s significant investment and talent acquisition underscore the competitive landscape in AI development, with both companies pursuing cutting-edge technologies.

Meta is investing $14.3 billion in Scale and has poached the startup’s chief executive officer, Alexandr Wang, for a new so-called 'superintelligence' unit, focused on building a more powerful, and hypothetical, form of AI software.

OpenAI’s pivot away from Scale AI signals a move towards more sophisticated AI architectures, emphasizing reasoning and autonomous task execution over traditional data-labeling partnerships.

This strategic realignment highlights the evolving priorities in AI research and development as companies race to build next-generation artificial intelligence systems.

OpenAI is ending its collaboration with data-labeling startup Scale AI, shortly after Meta invested $14.3 billion in Scale and hired its CEO. OpenAI is now focusing on developing advanced AI models that mimic human reasoning and perform tasks with minimal user input.