Labelbox was founded in 2018 by Manu Sharma, Brian Rieger, and Dan Rasmuson in San Francisco. The company recognized early that the quality of training data is often more important than model architecture, and built a platform to manage the full data lifecycle for AI.
Labelbox raised over $190 million, including a $79 million Series D led by SoftBank Vision Fund 2 in 2022. The company’s growth tracks the broader industry realization that data quality is the biggest bottleneck in production AI.
The platform covers data labeling (with human annotators, model-assisted labeling, and automated pipelines), data curation (finding and fixing issues in datasets), and model diagnostics (understanding where models fail and what data would improve them). Labelbox supports images, video, text, documents, geospatial data, and 3D point clouds.
Labelbox serves over 4,000 organizations, including enterprises in automotive (autonomous driving), defense, healthcare, and retail. Their Boost service provides managed labeling teams for organizations that need high-quality annotations at scale.
The company integrates with major ML platforms and model training pipelines, positioning itself as the data layer that sits alongside model development tools. Labelbox employs around 250 people and competes with Scale AI and V7 in the training data platform market. Their focus on the iterative nature of data improvement — not just one-time labeling — resonates with ML teams working on production systems.
