Why Scale AI matters
Meta paid $14.3 billion for a 49% non-voting stake in Scale AI in June 2025 and hired founder Alexandr Wang, after which Google, OpenAI and xAI cut or reduced contracts. CEO Jason Droege, who took over in June 2025, told Axios in May 2026 the company had a strong second half of 2025 and is now profitable with more than 1,300 employees. Scale is repositioning around reliability testing and enterprise agent deployment.
The numberMeta bought a 49% stake for $14.3B in June 2025; Scale said it was profitable as of May 2026
On The All Things AI 100
Listed under Data and infrastructure. Databases, data pipelines, and the plumbing under AI applications.
Quick answers
- What does Scale AI do?
- Supplies training data, evaluation and deployment services for AI models to labs, enterprises and governments. Meta paid $14.3 billion for a 49% non-voting stake in Scale AI in June 2025 and hired founder Alexandr Wang, after which Google, OpenAI and xAI cut or reduced contracts.
- What AI capabilities does Scale AI offer?
- Scale AI provides training data, evaluation, and deployment services for organizations that build AI models. Its Data Engine collects and labels data and supports human feedback for model training. Scale Evaluation measures model quality, safety, and performance. The GenAI Platform helps enterprises customize and run models on their data. Donovan gives government users document search, summarization, and analysis inside secure workflows.
- Why is Scale AI on The All Things AI 100?
- Scale AI is listed under Data and infrastructure. Databases, data pipelines, and the plumbing under AI applications. Meta bought a 49% stake for $14.3B in June 2025; Scale said it was profitable as of May 2026.
- Where is Scale AI based, and who owns it?
- Scale AI is based in San Francisco, CA, was founded in 2016 and is private; Meta holds a 49% non-voting stake (June 2025).