OpenWALDO maps 124B training tokens

OpenWALDO’s public training-data corpus now indexes 124 billion reference tokens across 75.1 million documents, organized into 20 corpora and 1,051 shards. The OpenWALDO project treats data like inspectable open-source input, while Techstrong’s launch report describes its AI Bill of Materials for carrying licenses and lineage into model releases. For builders, the filing-away point is that dataset provenance is becoming release infrastructure rather than a legal appendix; the open-weight coalition’s scale bet is the adjacent signal.