FinTech Giants Race to Build AI Foundation Models on Proprietary Data

Payments and financial data platforms including Mastercard and Plaid are turning billions of anonymized transaction records into training resources for generative AI. Their proprietary data covers spending, payments and account-linking behavior, potentially enabling foundation models tailored more closely to financial use cases and affecting cybersecurity defenses, loyalty programs and personalized services.

As of July 20, 2026, companies in the sector had begun developing AI foundation models based on proprietary data, shifting the focus of competition from general-purpose models to "intelligent finance" applications. The available information does not disclose investment amounts, model launch dates or the full number of transactions, stating only that the models are being trained on billions of anonymized transactions.