Where your AI runs is becoming part of what you sell
For Gulf companies, data residency used to be a line in a security questionnaire. Saudi enforcement, new in-Kingdom cloud regions and DIFC’s AI rules are turning it into a buying question, and data-first, private AI is how you answer it.
Here is a question worth putting to any leadership team in the region. When an employee pastes a client contract into an AI tool, where does that contract go?
In most companies, nobody can answer within a minute. Within a couple of years, clients will expect the answer in writing.
From checkbox to buying question
Data residency has lived in a corner of the IT department: a setting in a cloud console, a line near the end of a vendor questionnaire. Three forces are moving it to the front of commercial conversations in the UAE and Saudi Arabia.
The rules now have teeth. Saudi Arabia’s Personal Data Protection Law has been fully enforceable since 14 September 2024. In early 2026, the Saudi Data and AI Authority announced that its committees had issued 48 decisions against organisations in breach over the previous year. Fines reach SAR 5 million and can be doubled for repeat violations, and the law reaches anyone processing the personal data of people in the Kingdom, wherever that organisation is based. Personal data may leave the Kingdom only on the terms its transfer regulations set out.
In Dubai, the DIFC’s Regulation 10 sets rules for personal data processed by autonomous and semi-autonomous systems: a notice to users at first use, a register of how each system is used, and built-in triggers for human intervention where processing could be unfair. Certification requirements are expected to follow. The direction is clear: AI that touches personal data will have to show its working.
The infrastructure is arriving. Microsoft’s Saudi Arabia East cloud region opens to customers this month, with three availability zones in the Eastern Province and local data residency for eligible workloads. Amazon Web Services has said its US$5.3 billion Saudi region will open in December. The old objection, that the AI a company needs only runs abroad, gets weaker every quarter.
Buyers are rethinking who they depend on. In PwC’s latest CEO survey, 32% of Middle East chief executives said they plan to reduce reliance on technology providers based in countries they consider less trustworthy, against 21% globally. The same report notes that the Middle East now ranks second in the world for the average cost of a data breach, at more than US$7 million per incident.
Put the three together and residency stops being only a compliance cost. It becomes part of the offer, like delivery time or a warranty.
What “private AI” has to mean
“Private AI” is fast becoming a marketing phrase, so it is worth being precise. Four tests separate a real answer from a reassuring one.
- The data stays in the region. It is stored in the UAE or Saudi Arabia, in the company’s own cloud account or in an environment run for that company alone.
- The data never trains anyone else’s model. Not the vendor’s, and not a model provider’s.
- The AI sees what each employee may see, and nothing more. Permissions follow the organisation, not the history of the shared drive.
- Every output can be traced to its source. A document and a page for every figure, and a plain “not in your files” when the answer isn’t there.
Ask any vendor, including us, to prove all four in writing. Anything less is a promise, not a design.
The commercial argument
This is where it gets interesting for anyone who sells. Picture two firms pitching for the same contract. Both use AI to work faster. One of them can also say: your files never leave the UAE, they never train a model, and every figure in our work traces back to your own documents.
Only one of them has answered the question the client’s legal team was about to ask.
In any business that handles client information, that answer will increasingly decide shortlists. In procurement, the supplier who can complete the data-handling section in one clear page moves faster than the one who needs three weeks and a lawyer.
Our forecast: by 2028, “Where does your AI run, and on whose data?” will sit in standard procurement questionnaires across the Gulf, next to the questions on cybersecurity. Companies that answer it well will win work. Companies that can’t will lose it quietly, without ever being told why.
Private and data-first belong together
One without the other doesn’t hold. Well-governed data sent to an AI abroad is still a risk. A private engine running on scattered, unowned data is still unreliable: in the region, compliant, and wrong.
That is why Unknot Technology Labs Ltd works in that order. We govern a company’s data first, then build a private AI engine on it. Your data is stored only in the UAE or Saudi Arabia, in your own cloud or in an environment we run for your company alone, and never trains AI.
The infrastructure question is being settled this year. The commercial one is just starting. Companies that can say where their AI runs, in one sentence, will find that sentence opens doors.
Ahmed A. H. Hamid is the founder of Unknot Technology Labs Ltd at the DIFC Innovation Hub in Dubai. He spent more than eight years in business development and strategy consulting with medium and large enterprises and professional-services firms in the UAE and Saudi Arabia.
Sources: Latham & Watkins, Active Enforcement of Saudi Arabia Privacy Regime (May 2026); Mayer Brown, AI Regulation in the DIFC (January 2026); Microsoft, Saudi Arabia East region available in November 2026 (August 2026); AGBI, Microsoft’s Saudi cloud region (September 2026); PwC, 29th Global CEO Survey: Middle East findings (January 2026).
See it on ten of your own documents.
A 45-minute session at your office. Bring contracts, policies or price lists; leave with answers, finished work and their sources.
We reply within one business day.