AI firms pivot to licensed book datasets as web-based training data reaches saturation

Technology companies are increasingly acquiring large-scale book repositories to sustain LLM development as high-quality public web data becomes exhausted. This shift highlights a strategic move toward proprietary, curated datasets to bypass copyright litigation and data quality constraints inherent in web scraping.