OpenAI and G42's Inception42 plan custom AI for big UAE organisations

What was reported

As Crypto Briefing reported, citing ADMO, OpenAI and Inception42 will offer enterprise AI solutions customised for large organisations in the United Arab Emirates. The focus is big institutions, not consumers and not small businesses.

The claim is narrow, and it is worth being clear about how narrow. ADMO is the only source for it. There is no contract value, no timeline and no list of clients. Nor is there a statement from either company explaining who does what.

The word "tailored" hints at customised deployments rather than the standard off-the-shelf ChatGPT product. The report does not say what exactly gets customised, whether that means the model, the data it connects to, the language it speaks or the rules it follows.

Who Inception42 is

Inception42 is the sovereign agentic AI arm of G42, the UAE AI company. "Agentic" means software that carries out multi-step tasks on its own rather than only answering questions. "Sovereign" means the technology is meant to stay under national control, which matters to governments worried about depending on foreign systems. The company serves UAE government and enterprise clients and is led by CEO Ashish Koshy.

The report lists a busy run of activity. In May 2026 it was part of a national-scale AI initiative under which it aims to have AI agents handle 50% of federal government operations. In July 2026 it announced a deeper link between its Catalyst platform and Microsoft 365 Copilot, framed as support for the country's AI goals.

That same month it launched Seraj, an Arabic-language model for business use. Seraj was built on OpenAI's GPT-4.1 and then adapted with Arabic datasets, so Inception42 and OpenAI already had a technical connection before this announcement.

In October 2026, at AI Everything Abu Dhabi, the company introduced Insight, a platform meant to help with decisions. According to the report, more than 30 boards across UAE government and business already use it.

OpenAI's own UAE footprint

OpenAI is not arriving from nothing. The report says it has been building its own presence in the country, mainly through Stargate UAE, a project that gathers several partners around AI infrastructure, meaning the data centres and computing power that AI runs on.

One named customer connected to that effort is Emirates NBD. The report presents the bank as an early user in the MENA region of OpenAI's sovereign AI setup. That gives OpenAI a direct line to big UAE institutions, which raises an obvious question: why add a partner?

A plausible answer, though this is our reading and not the report's, is that large organisations often want a local company that understands their rules, their language and their politics. A partner with existing government and enterprise relationships can supply that. The report does not confirm this reasoning.

Separately, our newsletter today covered a reported agreement among ADNOC, G42 and OpenAI for the energy sector. That is a different deal, but it points the same way: OpenAI is working with Gulf institutions on several fronts at once.

The Microsoft question

The report flags one relationship as the variable to watch. Inception42 has tied its Catalyst platform more closely to Microsoft 365 Copilot, and Microsoft remains a central route by which OpenAI's technology reaches businesses.

So a company that works closely with Microsoft is now also set to work with OpenAI directly. How those channels fit together is unclear. They could complement each other, or customers could end up choosing between overlapping options. Nothing in the report settles it.

What to look for next

Until more detail arrives, treat this as an intention rather than a finished product. The report itself names the pieces that would make it concrete:

• Named clients that have signed up.
• Deployment timelines.
• Contract terms.
• A statement from OpenAI or Inception42 on how responsibilities are divided.

The last item matters most for ordinary readers. When an AI system is customised by a partner and run on a model from another company, a mistake raises a practical question about who fixes it and who answers for it. Announcements rarely address this; contracts and case studies do.