Frequently Asked Questions

Sphere is the marketplace for AI companies to source data with legal clarity. We provide a platform where creators and data providers can offer their content, allowing AI companies to purchase and use it in perpetuity for training and grounding their models. This ensures high-quality, factual information is available while establishing clear attribution and ending IP disputes.

Sphere sells high-quality, rights-cleared data to all major AI labs and developers who wish to build advanced AI products using data not found on the public web. We bridge the gap between AI research and the physical world by providing synchronized multi-modal data—such as robotic telemetry, egocentric video, and industrial sensor logs—that is essential for achieving true AGI and training sophisticated humanoid systems. Our marketplace serves as the primary source for the 'dark data' required to push the boundaries of embodied AI and complex world modeling.

Buyers can access a wide array of high-fidelity, non-web data including synchronized multi-modal captures (RGB, depth, IMU, force/torque), egocentric robotic telemetry, industrial sensor logs, and proprietary creative content. We specialize in providing the 'dark data'—information that isn't crawlable by standard bots—that is critical for training world models and physical AI.

Transactions on Sphere are structured as licensing agreements. This allows the buyer to use the data in perpetuity to train and refine their AI models, while the original seller retains their underlying IP and copyright. This model provides the legal certainty AI labs require for model grounding while protecting the long-term rights of the data providers.

Sphere employs a rigorous verification process to ensure all sellers have authoritative ownership of the data they provide. This eliminates legal risks for buyers, as Sphere guarantees the provenance and rights-cleared status of every dataset. Additionally, we use automated and manual quality checks to ensure data integrity and technical compliance with modern AI training standards.

Yes. Sellers have the option to offer their data on an exclusive or non-exclusive basis. Exclusive deals allow a buyer to be the sole licensee of a specific dataset, providing a significant competitive advantage for training unique models. These preferences are clearly marked on each listing within the marketplace.

The Sphere marketplace operates on a demand-driven model. Buyers can signal specific data needs or 'bounties' to the community. Sphere then alerts relevant sellers and data providers about this demand, facilitating the rapid acquisition of targeted datasets that may not yet be available on the open market.

Anyone with high-quality, proprietary data can apply to be a seller—from individual creators to industrial enterprises. Sphere vets every applicant to ensure they actually own the data and that it meets our high standards for accuracy and technical utility. This ensures the marketplace remains a trusted environment for the world's leading AI labs.

Sphere's percentage is 20% of the net price.