Oil companies have spent years using AI to drill faster, predict equipment failures and squeeze more production from existing fields. The UAE’s state-run oil giant, ADNOC, is now finding out what the same technology can do when the problem isn’t efficiency, but a war that initially cut UAE oil exports by roughly two-thirds, from about 5.1 million barrels per day before the conflict to just 1.9 million bpd in March. Almost seven months later, ADNOC is operating through a very different export system that heavily relies on AI.
By mid-September, UAE crude exports had recovered to 3.236 million bpd, up from 2.886 million bpd in August and 2.871 million bpd in July, according to Kpler data cited by Reuters. ADNOC has pushed crude through the Habshan Fujairah pipeline, expanded its tanker operations and used ship-to-ship transfers in the Gulf of Oman to keep crude moving outside Hormuz. Reuters reported last month that ADNOC had also become a major buyer of discounted Iraqi crude, processing much of it at Ruwais while freeing more UAE crude for export.
At the same time, AIQ’s technology is feeding ADNOC real-time information across drilling, production and facilities that CEO Dennis Jol says went from convenient to crucial almost overnight.
ADNOC went into the war with AI already embedded quite heavily. Systems developed by AIQ can automatically adjust producing wells, predict failures in critical equipment and analyze reservoir conditions, while newer technology is designed to connect decisions across the operation. Speaking to Semafor, AIQ CEO Dennis Jol has described a system capable of determining which wells to shut or keep producing when pipelines become unavailable and rerouting flows around the disruption.
Related: G7 Moves to Release 100 Million Barrels to Counter Diesel Crisis
By June this year, AIQ had developed roughly 200 AI use cases at ADNOC. Its technology includes predictive maintenance, safety applications, subsurface modeling and systems capable of automatically shutting down equipment to prevent accidents. Jol said some projects that previously took weeks can now be completed in hours.
The AIQ’s RoboWell system, for example, continuously optimizes producing wells using real-time data and AI models, and has now been deployed across more than 500 wells, increasing production by 5%, while reducing well interventions by as much as 50%. For ADNOC, that means individual wells can be continuously adjusted as production requirements change rather than waiting for each intervention from an engineer. The same technology can identify operating conditions that increase the risk of trips or unplanned shutdowns, giving operators another way to manage production when the rest of the oil network is changing quickly.
AI is also being used to detect equipment problems before they hinder production. Neuron 5 continuously analyzes pressure, temperature and vibration data from compressors, valves, generators and other machinery to predict maintenance requirements. ADNOC’s initial deployment covered hundreds of pieces of equipment at its Northeast Bab field and Taweelah gas compression plant. Pilot results indicated the system could reduce unplanned shutdowns by 50% and extend planned maintenance intervals by 20%.
ADNOC subsequently expanded Neuron 5 considerably, and by the end of 2024, it was operating across 1,200 pieces of critical equipment, with full deployment across the company scheduled for completion by 2027. ADNOC has since said Neuron 5 has cut unplanned shutdowns by 50%.
And now, AIQ is testing that predictive capability outside the UAE. In Egypt, its technology identified an impending failure in an electrical submersible pump 45 days in advance, giving operators time to intervene before the equipment went offline.
ADNOC is also using AI in subsurface work. During a 90-day ENERGYai trial across two fields, one AI agent completed seismic interpretation 10 times faster than conventional workflows, while another produced well pressure predictions in 15 minutes. In March 2025, ADNOC awarded AIQ a three-year, $340 million contract to deploy ENERGYai across its upstream operations, with the rollout eventually set to cover more than 28 producing fields and thousands of wells.
Even before the war, ADNOC said more than 30 AI applications generated $500 million in additional value in 2023 through lower capital, operating and working capital costs and improved production.
At a Semafor event in June, Jol said AI could help ADNOC decide which wells to shut down, which to keep producing and how to reroute flows when pipelines become unavailable. “We could have decided which wells we’re going to turn off today, which ones we’re going to pump, which pipelines were not working, and be able to reroute everything,” he said.
AIQ is now trying to turn the technology developed inside ADNOC into a business other oil companies can use.
According to Semafor, AIQ’s next step is Genesis, an operating system designed to run large scale agentic AI across upstream and downstream operations. Six months of war have added urgency to the rollout, according to Semafor. Genesis is also model-agnostic, allowing it to operate without tying customers to a single underlying provider such as OpenAI or Anthropic.
AIQ is already testing its technology in Kuwait, India, Malaysia and Vietnam. Egypt is separately discussing the creation of an AIQ Egypt venture that would combine data available through the Egypt Upstream Gateway with AIQ’s technology. The two sides are also examining AI applications for hydraulic fracturing, horizontal drilling, well design and exploration.
AIQ has spent six years developing technology inside ADNOC, with access to producing assets and decades of proprietary operating data. It now has to persuade other producers to put the systems developed through that relationship into their own operations.
As AIQ pushes further into international markets, the U.S., Canada and the North Sea are all among its targets. The company is also looking at acquisitions to accelerate that expansion, with CEO Dennis Jol telling Reuters that AIQ has substantial cash available and that deploying it is “up front and center.” According to Semafor, the company has made its first hire in the UK, is considering Houston and has identified roughly 100 potential acquisition targets as it expands.
By Tom Kool for Oilprice.com