NVIDIA and Elith have released Nemotron-Personas-Finance-Japan, an open synthetic dataset intended to let developers build and test financial AI without using real customer records. It extends the existing Nemotron-Personas-Japan collection and contains 1 million records, each holding nine personas, for a total of 9 million personas written in natural Japanese. Each record has 50 fields covering demographics, household structure, regional distribution, asset holdings, personality traits and financial behaviour, and the dataset totals roughly 1.5 billion tokens. The dataset is built on official Japanese government statistics using NVIDIA's NeMo SDG-PGMs, a probabilistic graphical model framework, and NeMo Data Designer. Its design was shaped by feedback from financial institutions including Sompo Japan, Tokio Marine Holdings, Bank of Fukuoka, Mizuho Financial Group and Meiji Yasuda Life Insurance, as well as the Financial Data Utilization Association. It is released under the CC BY 4.0 licence, which permits commercial use, and is positioned for conversational agents, customer simulation and model evaluation, with integration into NVIDIA's NeMo User Sim framework. The authors say it is the first open persona dataset at million-record scale to pair official statistics with household finance and financial literacy data, a claim that has not been independently assessed. For Australian businesses and professionals, the direct impact is limited. The dataset reflects Japanese demographics and is written in Japanese, so it is mainly relevant to teams building financial products for Japan. Others may find the open licence and synthetic-data approach useful for testing and evaluation. Using synthetic rather than real customer records can reduce the personal information handled during development under the Privacy Act 1988, but synthetic data is not automatically exempt, so organisations should confirm outputs cannot be linked back to real individuals before relying on that position.