Prakhar Mehrotra, PayPal’s global head of AI, said industry has moved from an information era to an 'intelligence era' where AI generates data on demand and 'tokens'—the atomic units of language models—become central to computation. He argued firms must invest in AI factories and token strategies to produce reliable, generated outputs. For academic programs, the shift changes how institutions teach data literacy and model governance: token economics matters for fines‑grained privacy, cost modeling and reproducibility. 'Tokens' here are the discrete text units models consume and produce—both a training input and the generated output metric universities must account for in curriculum and research budgeting.