Abu Dhabi signs four OpenAI deals for government and energy

Machine readable rules for officials, OpenAI's Frontier platform at ADNOC, AI for Mubadala's investment teams and a five year adoption deal.

On November 2, Abu Dhabi signed four agreements with OpenAI after a meeting between Crown Prince Sheikh Khaled bin Mohamed and Sam Altman. They cover machine readable government rules, OpenAI’s Frontier platform at the energy company ADNOC, AI for investment analysis at Mubadala, and a five year adoption deal.

Key facts

Abu Dhabi signed four agreements with OpenAI on November 2, 2026, witnessed by Crown Prince Sheikh Khaled bin Mohamed after he met OpenAI CEO Sam Altman.
A Machine Readable Government initiative turns selected rules into a format AI can read, so it can help officials with routine checks such as permits and certificates.
ADNOC becomes the first energy company to deploy OpenAI’s Frontier platform, with G42’s subsidiary Inception42.
Mubadala, Inception42 and OpenAI will use AI for investment analysis and risk assessment, and a separate five year deal covers AI adoption in government, business and education.
The government and ADNOC projects run OpenAI’s GPT models on inference infrastructure inside the UAE.

What happened

The Abu Dhabi Media Office says the Crown Prince, who also chairs the Abu Dhabi Executive Council, discussed with Altman how advanced AI can speed up government work and innovation in key sectors, and then witnessed the signing of four agreements.

The first, between the Abu Dhabi government and OpenAI, launches what it calls a Machine Readable Government. A defined set of policies and regulations is converted into a format that AI systems can read, so they can help officials with routine compliance checks. Every assessment can be traced back to the rule behind it, and officials keep oversight of and accountability for the final decision. The Department of Government Enablement leads the work, and a first pilot on food imports is already running with the Abu Dhabi Agriculture and Food Safety Authority.

The second makes ADNOC the first energy company to deploy OpenAI’s Frontier technology, together with Inception42, to support operational, investment and commercial decisions. The third brings AI to Mubadala for investment analysis, finding opportunities and assessing risk. The fourth is a five year agreement between Inception42 and OpenAI to promote AI across government, enterprise and education in the UAE, with joint enterprise services. The National notes that Frontier, launched in February, is OpenAI’s platform for building and managing AI agents in companies.

Why it matters

Most talk about AI in government is about chatbots that answer citizens’ questions. This deal goes one step deeper: the rules themselves are rewritten so a model can check a request against them, while a person still signs off. If it works on food imports, the same approach could reach other permits and certificates, and other governments will be watching.

There is a practical lesson for everyone else too. AI is far more reliable when the rules it must follow are written clearly, in one place, with nothing left to guess. That is true for a ministry and just as true for a small business with a returns policy, a school with a homework policy, or a team with an expense rule. The traceability point matters as well: each answer points to the rule behind it, which is exactly what you should ask from any chatbot that checks something for you.

The deals also show how much OpenAI’s business now depends on large organisations adopting agents through Frontier, not only on people paying for ChatGPT.

How to try it

You do not need a government programme to make your own rules easier for AI, and for people, to follow. Our free Rewrite text more clearly prompt turns a confusing paragraph into something shorter and clearer, keeps your meaning and shows what changed. Fill in the fields like this:

Text to rewrite: paste one rule you use often, for example your team’s travel expense policy or your shop’s returns policy.
Who will read it: “New staff members and an AI assistant that checks requests against this rule”
Tone: “Formal”

When you get the answer, check three things. First, that every condition from the original, such as limits, deadlines and exceptions, is still there; if one is missing, point it out and ask for it back. Second, that each condition is now a separate sentence or bullet, so a person or a model can check them one by one. Third, that the rewrite did not add rules that were not in your text. Once you are happy with it, you can paste the clear version into a chatbot together with a real request and ask it to say which rule applies and why.

Try it

Rewrite text more clearly

Frequently asked questions

What is Abu Dhabi's Machine Readable Government with OpenAI?

It is an initiative led by the Department of Government Enablement that converts a defined set of policies and regulations into a format AI can read. AI then helps officials with routine compliance checks, each one traceable to its rule, while officials keep the final decision. The first pilot covers food imports.

What did ADNOC agree with OpenAI?

ADNOC, with G42's subsidiary Inception42, agreed to deploy OpenAI's Frontier platform across its operations to support operational, investment and commercial decisions. It is described as the first energy company to do so, and it runs OpenAI's GPT models on infrastructure located in the UAE.

Does this change anything for ChatGPT users?

Not directly. The agreements concern Abu Dhabi's government, ADNOC, Mubadala and enterprise services through Inception42. For everyday users, the useful idea is the method: write rules clearly before asking AI to apply them, and ask it to show which rule supports each answer.

Sources