This article draws on work by the TUM Think Tank, Stellenbosch University, and members of the Global Network of Internet and Society Centers on regulatory sandboxes for digital public infrastructure. Insights were also generated at a workshop on “Regulatory Sandboxes for DPI in South Africa” and a side event at the 2025 UNESCO Global Forum on the Ethics of Artificial Intelligence.

Regulatory sandboxes (sandboxes) are temporary, controlled, and supervised environments in which new technologies, business models, or regulatory approaches can be tested before being subject to the full force of existing rules.(1) They typically include limited regulatory exemptions, supervisory oversight, and evidence generation for future policy decisions.(2) Their use has expanded significantly across financial services, telecommunications, artificial intelligence, and digital infrastructure, and they are increasingly treated as standard instruments in the governance toolkit for emerging technologies.(3)