According to Mark Cuban on Twitter, the future of AI large language model (LLM) training will rely on exclusive access to proprietary data, as intellectual property holders increasingly sell their data only to the highest bidders. This fragmentation of AI training sources could impact blockchain-based data marketplaces and reinforce the demand for decentralized data solutions in the crypto sector, as traders seek new ways to monetize and access exclusive datasets (source: Mark Cuban, Twitter, June 11, 2025).