Why trust Greenbot
Google is walking away from a $200 million partnership with Scale AI after rival Meta grabbed a 49% stake in the data-labeling startup, sources told Reuters on June 14.
The search giant had planned to spend that money this year on human-annotated training data crucial for developing its Gemini artificial intelligence (AI) models. But Meta’s $14.3 billion investment changed everything. Scale AI’s CEO Alexandr Wang is now joining Meta to lead its “superintelligence” efforts.
Microsoft and Elon Musk’s xAI are also heading for the exits. OpenAI pulled back months ago but maintains Scale as “one of many vendors,” its CFO confirmed.
The exodus stems from a simple fear: sharing secrets with the enemy. AI companies worry that working with Scale could expose their research priorities and technical blueprints to Meta through the startup’s operations.
“Companies that compete with Meta in developing cutting-edge AI models are concerned that doing business with Scale could expose their research priorities and road map to a rival,” Reuters reported, citing five sources.
Scale AI’s business model depends on a network of highly trained human annotators. PhD-level scientists and historians label complex datasets through a process called Reinforcement Learning from Human Feedback. These expert annotations can cost up to $100 each.
The startup generated $870 million last year, with Google contributing $150 million as its largest customer. Meta’s investment doubled Scale’s valuation from $14 billion to $29 billion practically overnight. The deal transforms what was once a neutral vendor into an extension of one of the AI race’s biggest players.
Competitors are already capitalizing on the chaos. Labelbox CEO Manu Sharma told Reuters his company will “probably generate hundreds of millions of new revenue” by year’s end from fleeing customers.
“Our demand has tripled overnight after the news,” said Garrett Lord, CEO of rival firm Handshake.
Jonathan Siddharth, who runs competing startup Turing, captured the moment’s significance. “Leading AI labs are realizing neutrality is no longer optional, it’s essential,” he said.
Google had been diversifying its data service providers for over a year, three sources said. Meta’s move accelerated those plans. The company is now shifting workloads to competitors like Labelbox and Handshake.’
The fracture highlights the growing importance of specialized human expertise in AI development. While the race once focused on securing computer chips and electricity, it now extends to capturing the world’s smartest annotators.
Scale AI declined to comment on its relationship with Google. A company spokesperson told TechCrunch that Scale’s business remains strong, spanning major companies and governments, and emphasized its commitment to protecting customer data.
The startup continues to serve self-driving car companies and U.S. government agencies, which are expected to remain its clients. But losing generative AI customers like Google threatens Scale’s core business model.
Meta’s strategic gambit sacrifices Scale AI’s vendor neutrality to absorb top talent for its own AI ambitions. The move signals that in the race for artificial intelligence supremacy, even the supply chain has become a battlefield.
