DataMasque raises $5.6m for safer AI data privacy tools

DataMasque has raised $5.6 million after pitching privacy-safe data access as a practical choke point for enterprise AI projects. The Australian de-identification startup said Wavemaker Partners led the round, with OIF Ventures and Icehouse Ventures returning, SmartCompany reported.

Its pitch is practical. Development teams need records that behave like real customer files before they can train, test or evaluate models. Privacy teams still have to keep names, account details and other identifiers out of test environments and analytics tools. DataMasque says its software masks that personal information while keeping the relationships inside a dataset intact.

Grant de Leeuw, DataMasque’s co-founder and chief executive, told SmartCompany the funding was tied to that access problem.

“Enterprise AI projects are increasingly blocked by data access. This funding validates the need for synthetically identical customer data to feed into AI models or use to evaluate, test, and train AI,”

The company said annual recurring revenue had grown six-fold, while 95 per cent of revenue now comes from production deployments. That production figure matters in a market where enterprise AI pilots can stall long before they reach operational systems.

DataMasque also said one customer had more than 100 AI initiatives blocked because teams could not safely access suitable data. For a chief information officer, that turns de-identification from a legal control into an operational dependency.

AI adoption runs through test data

Before an enterprise can train or evaluate a model, engineers usually need a dataset that reflects actual customer behaviour. Security and privacy teams then ask where that data will go, who can see it and whether sensitive fields will survive the trip. If masked data loses too much structure, the tests can point teams in the wrong direction. If the data cannot move at all, the project stops even earlier.

DataMasque calls the resulting datasets “synthetically identical”. In practice, that means replacing personal information while keeping the patterns that matter for analytics and software testing. In a DataMasque post about its strategic collaboration agreement with Amazon Web Services, Wavemaker’s Paul Santos said enterprises needed “a way to protect sensitive information without compromising datasets used for testing and analytics”.

The round puts DataMasque in a busy slice of the Australian startup market. Enterprise AI spending has lifted demand for infrastructure below the chatbot layer: data access, security review, cloud deployment and audit controls. Those systems often decide whether a bank, insurer or retailer can move from trial to live deployment.

DataMasque previously raised $2.3 million in seed funding in 2023. The new round is larger, but the sharper claim is the shift to production usage. If the reported revenue mix holds, customers are buying de-identification as an operating layer for AI and analytics programmes, not as a one-off privacy clean-up.

For Australian enterprise technology, that gives the funding round a clearer read than a routine capital raise. The AI market already has plenty of model vendors and app-layer tools. DataMasque is arguing that the blocker sits closer to the database: sensitive records cannot help models improve unless companies can move a faithful version of that data safely.