Licensing Wikipedia to Fund the Commons

Wikipedia’s new enterprise contracts with Amazon, Microsoft, Meta, Perplexity and Mistral show a turning point: public, volunteer‑maintained knowledge platforms are beginning to sell structured access to AI developers at scale to cover server costs and deter indiscriminate scraping. This creates a practical business model for sustaining public goods while forcing AI firms to internalize training‑data costs. — If replicated, pay‑to‑train deals will reshape the economics of AI training data, set precedence for other public and cultural datasets, and force policymakers to decide how public knowledge should be priced, governed, or subsidized.