Banks want the benefits of cloud and AI. Regulators, however, care a great deal about where customer data sits. United Overseas Bank (UOB) says it has found a way to satisfy both sides, as it has signed a Memorandum of Understanding with Amazon Web Services (AWS) to support its hybrid cloud strategy.
Under the agreement, the two companies will build a dedicated cloud environment in Singapore. This infrastructure model lets UOB systems and data run across the bank’s on-premises infrastructure and the cloud. In practice, it works as an isolated cloud extension that UOB can deploy in specific countries or jurisdictions to meet local regulatory requirements. At the same time, the bank keeps access to AWS tools, including AI and data analytics.
UOB already runs an Enterprise GenAI Platform. Now, the bank plans to explore the capabilities of Amazon Bedrock further.
Lawrence Goh, head of group technology and operations at UOB, tied the move to the bank’s AI ambitions. He said AI can help the bank become more adaptive, intelligent and responsive. However, he noted that this requires a secure and resilient technology foundation, strong governance and the right partners. Goh also sees room to widen the collaboration over time and explore other areas with the wider Amazon group.
Jaime Vallés, vice president and managing director for APAC at AWS, said the collaboration aims to improve everyday financial experiences for millions of people across the region. In addition, he said AWS will give UOB training programs that build AI fluency and cloud skills across its workforce, while customer data stays secure.
The agreement also fits UOB’s wider investment in technology. In May 2026, the bank opened a global technology and innovation center in Singapore’s Punggol Digital District, after a $500 million investment.
The isolated-extension design may matter beyond Singapore. UOB, which dates back to 1935, runs banking subsidiaries in China, Indonesia, Malaysia, Thailand and Vietnam, and each market has its own data rules. Therefore, a model that UOB can place within a given jurisdiction offers a practical way to expand AI work without moving sensitive records across borders.
