Scale AI
Scale AI develops reliable AI systems for the world's most important decisions. It provides high-quality training and evaluation data plus full-stack technologies that help leading AI labs, enterprises, and governments build, deploy, and oversee AI applications.
At a Glance
- Frontier AI and foundation-model developers
- Enterprise technology, financial services, healthcare, media, and retail companies
- Autonomous-vehicle, robotics, and transportation companies
- U.S. federal, defense, intelligence, and other government agencies
- +1 more
AI Tools by Scale AI
(2)
AgentEnv Framework
Open Source RL Environment Framework
Scale AI
AI Data Labeling and Evaluation
Discussions
No discussions yet
Be the first to start a discussion about Scale AI
Latest News
To Steer the AI Frontier, Washington Must Build Its Testing Power
Scale AI appointed Francis deSouza as CEO, effective August 10, 2026
Scale AI joined the U.S. Department of Energy Genesis Mission Consortium
Scale AI expanded its Pentagon/CDAO enterprise agreement from a $100 million ceiling to $500 million
Products & Services
Managed data and annotation platform for image, video, text, and 3D sensor-fusion data. It combines machine-assisted labeling with human expertise for data generation, RLHF, red teaming, and model evaluation.
Enterprise platform to connect proprietary data, build and execute generative-AI applications and agents, evaluate and monitor them, and convert usage and human feedback into continuous improvements.
Expert-led model testing, red teaming, benchmarking, leaderboards, and safety/alignment research across coding, math, languages, vision-language, agentic tool use, instruction following, and adversarial robustness.
Secure, model-agnostic platform for customizing, evaluating, and deploying specialized AI agents for defense, intelligence, and other public-sector workflows, including document analysis, knowledge-base retrieval, planning, and decision support.
Market Position
Scale positions itself as a full-stack AI infrastructure and applications company spanning data generation, annotation, evaluation, and deployment, rather than only a labeling vendor. Its differentiation is the combination of large-scale human/expert data operations, proprietary evaluation and safety research, enterprise agent infrastructure, and secure public-sector deployment; competitors include Appen, TELUS International AI Data Solutions, Labelbox, Surge AI, Sama, Snorkel AI, and in parts of the platform market Databricks, Palantir, and other agent/evaluation platforms.
Leadership
Founders
Alexandr Wang
Worked as an engineer at fintech company Addepar and at Quora; he co-founded Scale AI and served as its CEO before joining Meta's AI efforts in 2025. He remains a Scale board director.
Lucy Guo
Worked at Quora as a product designer with a software-development background; she co-founded Scale AI with Alexandr Wang and later left the company while retaining equity.
Executive Team
Francis deSouza
Chief Executive Officer
Former Google Cloud Chief Operating Officer; appointed Scale AI CEO effective August 10, 2026. Previously served as CEO of Illumina and co-founded and led software company Symantec.
Board of Directors
Founding Story
Wang and Guo started Scale in 2016 after seeing that machine-learning teams were bottlenecked by the lack of high-quality labeled data. Their initial vision was an API for human intelligence: customers would send raw data to Scale, which combined machine learning with human review to return production-ready training data and accelerate AI development.
Business Model
Revenue Model
B2B and government sales through customized enterprise contracts for data production, annotation, evaluations, software-platform access, agent deployment, and professional services; the data/API business is usage-based while platform and public-sector engagements are generally contracted.
Pricing Tiers
Designed for experimentation and smaller projects; usage-based labeling and annotation charges apply beyond included usage.
Custom volumes, security/compliance, dedicated support, managed data and platform services, and enterprise agreements.
Target Markets
- Frontier AI and foundation-model developers
- Enterprise technology, financial services, healthcare, media, and retail companies
- Autonomous-vehicle, robotics, and transportation companies
- U.S. federal, defense, intelligence, and other government agencies
- International governments and public-sector organizations
- Training and post-training of frontier and enterprise AI models
- LLM and agent evaluation, safety testing, and red teaming
- Autonomous-vehicle perception and sensor-fusion systems
- Enterprise document processing, search, knowledge retrieval, and workflow automation
- Healthcare and financial-services AI
- Defense and intelligence analysis, planning, logistics, and decision support
- OpenAI
- Meta
- NVIDIA
- Toyota