Troveo, a provider of licensed real-world data for AI training, has expanded its platform into five new data categories and announced over $20 million in payouts to content owners. The company now offers training-ready datasets spanning video, audio, text, enterprise workflows, gaming and robotics. The expansion addresses a key bottleneck in developing frontier AI models: access to high-quality, non-public training data. Troveo's library includes over 8 million hours of licensed video footage, 4 million hours of multi-channel audio, billions of words from publishers, and real-world business data from enterprise environments. All datasets are sourced and licensed from content owners, providing legally defensible data pipelines. The company works with thousands of content owners and serves major AI labs and technology companies globally.

Insider Brief Troveo launched from stealth with $4.5 million in seed funding led by Seven Seven Six, providing a platform for content owners, like