Google is backing out of a key AI partnership as Meta steps into the mix. According to five insider sources, Google is preparing to sever ties with Scale AI – a major data-labeling company it has worked with extensively.
In fact, Reuters reports that Google was Scale AI’s largest client.
So what exactly does a data-labeling company do? In short, they help artificial intelligence learn. Employees meticulously tag or label images, videos, or text to help AI systems understand what they’re seeing or reading. For example, tagging a photo with ‘dog’ or marking a sentence as ‘sarcastic’ provides context that the AI uses to improve its recognition and understanding. It’s like giving AI a teacher to guide its learning.
Until recently, Google planned to pay around $200 million in 2025 to Scale AI to help refine its Gemini model. But now those plans appear to be on hold – if not scrapped entirely – because Meta is buying a 49% stake in Scale AI.
This sudden move by Meta, the tech titan behind Facebook, Instagram, and WhatsApp, has raised eyebrows across the AI industry. Companies like Microsoft, OpenAI, and Elon Musk’s xAI are reportedly spooked. Their concern? That Meta could gain indirect access to proprietary training data or insights via Scale AI – a potentially huge competitive risk in the AI arms race.
As a result, other AI firms are looking elsewhere. Competitors like Labelbox and Handshake are reportedly getting a surge of interest, and some organizations are even building in-house labeling teams to maintain full data privacy.
While Scale AI may lose some key clients, the company is unlikely to crash and burn. But there’s no denying that Meta’s surprise investment has disrupted the delicate balance of trust in the AI ecosystem.