Scale AI is the data infrastructure backbone of the modern AI industry, providing high-quality labeled training data, evaluation tools, and model customization services to companies building frontier AI. Founded in 2016 by Alexandr Wang (then 19) and Lucy Guo out of Y Combinator's S16 batch, Scale tackled the unglamorous but critical bottleneck of AI development: the need for massive volumes of accurately annotated data spanning images, video, text, LiDAR, and sensor fusion. What began as labeling for self-driving car companies grew into the central data pipeline behind LLMs from OpenAI, Meta, Microsoft, and the U.S. Department of Defense.
The platform combines a global human workforce (managed through Remotasks and Outlier) with proprietary machine learning tooling to produce expert-grade datasets, RLHF feedback, red-teaming, and model evaluations. Core products include Scale Data Engine for training data, Scale GenAI Platform for enterprise LLM deployment, Scale Donovan for defense and intelligence applications, and Scale Evaluation for benchmarking model performance. The company also runs the SEAL leaderboards, an independent third-party evaluation framework for frontier models.
Scale raised over $1.6B across multiple rounds, reaching a $13.8B valuation in 2024 led by Accel. In June 2025, Meta took a 49% non-voting stake worth approximately $14.3B, valuing Scale near $29B, and hired founder Alexandr Wang to lead Meta Superintelligence Labs. Jason Droege became CEO. Scale serves OpenAI, Microsoft, Toyota, the U.S. Army, and most major frontier labs, generating an estimated $870M+ in 2024 revenue.