Coordinating model agents in the energy sector: ADNOC ties reservoir and operations planning to G42 infrastructure
Abu Dhabi National Oil Company (ADNOC) signed a multi-year partnership agreement with OpenAI and G42, becoming the first energy company to use the OpenAI platform to coordinate swarms of intelligent agents across complex operational and commercial processes. The agreement was signed during the Abu Dhabi International Petroleum Exhibition & Conference (ADIPEC), in the presence of Sheikh Khaled bin Mohammed bin Zayed Al Nahyan, Crown Prince of Abu Dhabi and Chairman of the Executive Council of the Emirate of Abu Dhabi, to launch a new phase in which AI applications move from individual tasks to full-scale industrial deployment.
The partnership’s first phase focuses on two clear technical tracks. The first track aims to improve field development plans by integrating data from thousands of wells, multiple reservoirs and surface facilities with investment costs and commercial requirements. This system uses ADNOC’s existing ENERGYai platform together with OpenAI’s advanced models to craft integrated field plans as a single operating unit rather than studying reservoirs and facilities in isolation.
The second track unifies data flow across production, processing, refining, transportation, sales and commercial trading joints.In this context, the agent networks forecast oil supply and demand levels, price trajectories and profit margins, giving the company a continuous ability to measure the commercial and operational impact of its decisions and to extract the maximum possible value from each barrel before it leaves the ground.
The partnership relies on the OpenAI Frontier framework to develop these joint systems by merging advanced models with ADNOC’s industrial data and accumulated expertise. The technologies are hosted entirely on G42’s sovereign AI infrastructure in the United Arab Emirates, under full oversight by ADNOC staff to protect data and sensitive facilities, while keeping pivotal decisions in the hands of human experts. The agreement also includes creating a shared knowledge base to document operational expertise and train staff, alongside enhancing cybersecurity through the Daybreak Blue initiative to detect vulnerabilities and protect vital energy assets.
The alliance crowns a path ADNOC began by deploying disparate technical solutions in its facilities, such as the SMARTi occupational safety system, the RoboWell remote autonomous well-operation system, and the AiPSO platform developed with SLB and Cognite to improve reservoir production. Moving to OpenAI agents represents a leap that integrates those standalone tools into a single computing ecosystem that manages the entire workflow.
This shift imposes a new professional and regulatory equation on technology and operations teams across the Gulf and the Arab region.Actual implementation no longer stops at trying software assistants or buying off-the-shelf subscriptions; it now requires connecting live industrial data streams directly to inference engines, with companies obliged to run models on locally sovereign data centres to ensure sensitive energy data do not leak. If your organization manages large operational or industrial assets, your next step will require upskilling systems engineers to lead agent integration and govern their operational decisions, avoiding any uncalculated error that could affect physical infrastructure in the field.