Three things this table tells you.
1. The buyers are the largest companies on earth, and they are not stopping. Every frontier lab now runs a formal buying program for licensed data. OpenAI operates a public data partnerships program; Google negotiates content and data licences directly. Estimates put total lab spending on training data above US$10B a year and rising, because quality data, not compute, has become the binding constraint on model progress.
2. Nobody in this table is a data company. A newspaper group, a forum, a stock photo library, three academic publishers. Each was sitting on years of accumulated content and operational history that suddenly acquired a market. The same shift is now reaching business operational data: how companies actually communicate, decide and execute is exactly what the next generation of AI needs and cannot find on the public internet.
3. Clean rights command the premium. Every deal above exists because scraping is no longer defensible. Buyers pay for licensed, warranted, consented data because it removes legal risk from their models. That is precisely what Australian supply offers by default, and why we operate here.
Your business is smaller than Reddit. Your data is also something Reddit does not have: the complete private record of how a real company operates. Different asset, same market.