OpenAI has decided to end its data service agreement with Scale AI. The move comes after Meta invested billions of dollars in Scale AI and added its founder Alexandr Wang to the board. OpenAI shared this update with Bloomberg on Wednesday.
In recent weeks, Meta revealed a major deal with Scale AI. Meta said it would work with the startup to train its own AI models. Soon after that news, OpenAI began moving away from Scale AI. A spokesperson for OpenAI said the company was already looking for other data partners to support its next generation of AI systems.
Why OpenAI chose new data partners
Scale AI has long helped OpenAI label and prepare data for model training. Scale AI uses human teams and custom software to clean data. This work helps make AI responses more accurate. OpenAI’s CFO had once signaled continued cooperation. Yet the Meta announcement changed that tone. The OpenAI spokesperson noted a need for providers who remain independent of big tech rivals.
Scale AI faces its own challenges
OpenAI is not the only company reconsidering Scale AI. Reuters reported that Google is also planning to end its deal with Scale AI. Other AI labs have reached out to rivals of Scale AI to secure neutral data services. This scramble hints at a wider industry shift. Many model builders now seek partners that do not have ties to direct competitors.
In response, Scale AI’s general counsel wrote that the Meta deal will not grant special access to user data. Scale AI pledged to keep customer information private and said Wang will not run day-to-day operations. Yet the largest clients have already begun to leave.
Scale AI’s new focus
On Wednesday, Scale AI’s interim CEO, Jason Droege, announced a new strategy. Scale AI will focus more on building custom AI applications for large firms and governments. The startup aims to use its own tools to deliver AI solutions rather than only offer data services. Droege said the company will double down on these application offerings.
Scale AI’s shift marks a big change from its roots in data labeling. The new path may open fresh opportunities. Yet it also comes at a time when its main customers seek out other firms to avoid conflicts of interest.
What this means for the AI industry
OpenAI’s split with Scale AI shows how high the stakes have risen in AI. Major model builders now guard their data supply chains carefully. They worry that rivals may gain unfair advantages. As AI grows more central to many products, control over data has become a key strategic asset.
OpenAI will now work with new partners to ensure its data remains neutral. Scale AI will use its skills to build tailored applications in place of pure data work. Both moves could reshape how AI systems get trained and deployed in the coming years.
The drama around Scale AI is still unfolding. It points to a new era where data services carry as much weight as the AI models themselves. As companies explore new alliances, the market for data and AI applications will keep shifting in unexpected ways.
