Emirates NBD will run OpenAI’s GPT models on infrastructure inside the UAE, the bank and OpenAI said on 19 October, a deployment they describe as the first by a bank in the Middle East and North Africa.

The bank plans to apply the models in four areas: customer engagement across its digital channels, risk and compliance analytics, automating knowledge-heavy workflows, and employee productivity. It will connect them to its internal systems, interfaces and enterprise data, and names risk monitoring and financial crime controls among the expected gains. Emirates NBD serves more than 25 million active customers in 13 countries and had assets of AED 1.3 trillion at the end of June, according to the companies’ announcement.

Miguel Rio-Tinto
Miguel Rio-Tinto

Miguel Rio-Tinto, the bank’s group chief digital and information officer, said:

“AI is not an experiment for us, it is a core capability.”

Farouk El Hamzawi, head of enterprise at OpenAI MENA, said the bank is “setting a benchmark” for how financial institutions can deploy AI securely.

Also, read Mashreq Pilots IBM’s AI Agents on the Integration Platform Behind Its Payments

What in-country inference means

Data residency covers where a customer’s information is stored. Inference residency covers where the model runs. OpenAI added the UAE to its inference residency programme on 12 August, Middle East AI News reported, alongside the US and Europe as the only other regions. At launch it supported only one model, GPT-5.2, and some features were unavailable. Authentication and analytics can still run outside the UAE.

Emirates NBD hasn’t said which models it will use. The 12 August launch named no bank customers.

The regulatory backdrop

The companies say in-country inference helps with data residency, privacy and operational resilience, which they describe as expectations under local regulation. They name no regulator or rule.

The Central Bank of the UAE launched a sovereign financial cloud with Core42 in February for licensed institutions, which we covered here.

Marie Chowdhry, a fintech partner at law firm Pinsent Masons, said of that cloud that “data residency, supervisory access and resilience obligations are becoming increasingly operationalised and standardised.”

The central bank’s guidance on AI says institutions relying on third-party or cloud providers should check the provider’s governance, security and data-protection practices, according to its rulebook.

What’s next

The announcement gives no go-live date or rollout schedule, and doesn’t say where the servers are hosted or whether they sit on the central bank’s sovereign cloud. It publishes no figures on cost, performance or the number of staff and customers affected. The claim of a first has no independent comparison behind it.