Labelbox

Proximal, founded in fall 2025 in San Francisco, describes itself as a data research lab that builds complex, long-horizon reinforcement learning environments and evaluation data for frontier AI labs and enterprises. It started with coding-agent training data and is expanding into drug design, chip design and legacy software modernization; it also released the open benchmark FrontierSWE.

Micro1, founded in 2022 in Palo Alto, California, helps AI companies source and manage human contractors for data labeling and model training.

XDOF, founded in 2024 in Menlo Park, USA, builds data collection tools, data pipelines, and annotation systems for robot training data, serving AI labs and robotics companies. It addresses the bottleneck of insufficient physical manipulation data for general-purpose robot training.

Perceptron (Perceptron Network) is a ~2024-founded decentralized AI-data-infrastructure startup with a UK/UAE/Israel team. It condenses global data collection into a single contributor mesh, letting AI companies commission specific, human-verified datasets on-demand via its data-questing platform, delivered in days by ~800K nodes and 300K+ daily contributors across 150+ countries at ~90% lower cost than centralized data vendors. It merged with the 700K-node BlockMesh network in June 2025 to form an end-to-end DePIN data pipeline with token incentives.

Abundant (YC F24, founded 2024, San Francisco) provides on-demand expert human operators that handle AI agent edge cases via API with auditable decision records, and builds agent simulation / reinforcement-learning environments and datasets for AI labs; it says it serves three of the six leading AI labs and several Fortune 500 companies.

为模型实验室提供工程人才与训练数据