Emirates NBD runs OpenAI GPT models inside UAE borders

Key Points

1. Emirates NBD says it is the first MENA bank running GPT models on UAE-based inference.
2. The bank reports 25 million active customers and 1.3 trillion dirhams in total assets.
3. Keeping inference onshore lets Gulf banks apply generative AI to regulated customer data.

The latest:

Emirates NBD said on 19 October 2026 it has become the first bank in the Middle East and North Africa to run OpenAI’s GPT models on inference infrastructure located inside the United Arab Emirates, under an enterprise partnership with OpenAI. The bank framed the arrangement as keeping model processing within national borders while connecting the systems to internal platforms, APIs and enterprise data.

Details:

• The deployment: According to Emirates NBD, the partnership places GPT inference on infrastructure hosted inside the UAE rather than routing requests abroad. The bank said the models are being connected to internal systems, programming interfaces and enterprise data, which is the step that makes regulated customer information usable by generative models.
• The use cases: The bank listed five areas: customer service across digital channels, risk and compliance analytics, automation of knowledge-intensive tasks, employee productivity, and the integration layer linking models to its own platforms. Emirates NBD did not set out which of these is live and which is planned.
• The scale: Emirates NBD reported more than 25 million active customers and a presence in 13 countries. Total assets stood at 1.3 trillion dirhams, roughly 354 billion dollars, as of 30 June 2026 — the figure that sets the size of the data estate now in scope.
• The network: The bank said it operates 1,425 branches and 4,948 ATMs, while 97% of its transactions already take place outside branches. That digital share is the channel where customer-facing generative models would reach the largest number of users.
• The bank’s framing: Miguel Rio-Tinto, group chief digital and information officer at Emirates NBD, said artificial intelligence is a core capability and central to the future of banking. The bank presented the onshore setup as an operating model rather than a single product launch.
• OpenAI’s position: Farouk Al Hamzawi, head of enterprise for OpenAI in the Middle East and North Africa, said the bank is “setting a benchmark for deploying AI safely and responsibly.” OpenAI described the work as an institutional partnership with the lender.
• The policy fit: Emirates NBD tied the deployment to the UAE National Artificial Intelligence Strategy 2031, the state programme that sets government targets for AI adoption across sectors, positioning the bank’s move as aligned with national policy rather than a standalone commercial decision.
• The timing: No rollout dates or phases were announced. Emirates NBD did not name the infrastructure provider hosting the inference layer, nor did it disclose the financial terms of the arrangement with OpenAI.

Background:

Gulf regulators generally require banks to keep sensitive customer data processed domestically, which has limited the use of large language models hosted overseas. Onshore inference is the technical arrangement that lets a bank apply such models to regulated data without moving it out of the country.

Between the lines:

The newsworthy element is the location of the inference, not the models themselves. With 97% of Emirates NBD transactions already outside branches and more than 25 million active customers, the constraint on deploying GPT models at that scale was jurisdictional rather than technical. Framing the step against the 2031 national strategy suggests the bank expects the onshore model to be read as a template other Gulf lenders can follow.

What’s next

Watch for Emirates NBD to name the first live customer-facing use case and disclose a rollout timeline, and for other Gulf lenders or regulators to respond to the onshore inference model.