Search for a command to run...
ADNOC connects 120+ rigs to one AI operations centre, with large reported gains and unanswered questions about evidence, control and resilience.
No comments yet. Be the first to comment.
Passkeys remove a common phishing target, but a weak reset flow can undo the security gain. Start with recovery, then add credentials without locking users out.
Saudi clearance and UAE Peppol exchange demand different country adapters, not two billing platforms.
July funding rose to $172.6 million, but debt and two Gulf markets accounted for most of the increase.
A mobile-friendly guide to cloud availability, residency, procurement and recovery across four Gulf markets.
Sultan Byte
16 posts
Articles addressing day-to-day development challenges
On this page
ADNOC has connected more than 120 onshore and offshore drilling rigs to an AI-enabled Real-Time Operations Center built with SLB. This is a live fleet deployment, not a lab trial or a roadmap.
The operating model matters more than the "AI" label. Live rig data enters one environment, automated dashboards replace several tools, and engineers get earlier warnings about potential problems. That makes ADNOC's deployment a practical test of whether industrial AI can improve decisions without weakening operational control.
ADNOC announced the deployment on 4 August 2026. The Real-Time Operations Center, or RTOC, runs across the ADNOC Drilling fleet and supports ADNOC Onshore and ADNOC Offshore.
SLB says its DrillOps software brings live drilling data into a unified environment. Business, asset and drilling teams can see activity across the fleet instead of moving between separate tools and reports. Automated dashboards and analytics turn that data into alerts and operational information.
That distinction matters. A chatbot added to an old workflow may save a few minutes. A shared operational layer changes who can see an issue, how quickly teams respond, and how many assets an engineer can supervise. It also creates a larger failure domain. Bad data, a weak model, or an incorrect alert can travel further when the system covers a whole fleet.
ADNOC has not disclosed which rigs are connected, the model architecture, false-alert rates, or how much authority the platform has over equipment. Oil & Gas Journal noted that details of the specific rigs were not released. The public description points to decision support and workflow automation rather than fully autonomous drilling.
ADNOC says the RTOC reduces engineering effort by 30 to 40 per cent and lets engineers support two to three times more rigs while maintaining oversight. Reporting cycles that took several days can now finish within hours. Analyses that once required a full day can be completed within minutes.
The company also reports that earlier detection can shorten incident response by four to 12 hours and help avoid one to two days of rig downtime. The National confirmed the fleet-wide deployment, while Rigzone and Offshore Engineer reported the same figures.
Those figures still come from ADNOC. None of the public reports provides a third-party audit, a baseline period, a sample size, or a definition of "engineering effort." A 40 per cent reduction in time spent preparing reports is different from a 40 per cent reduction in total engineering work. Avoided downtime is harder to verify because it depends on a counterfactual: what would have happened without the warning?
Sources: ADNOC and SLB, 4 August 2026. Performance figures are company-reported and were not independently audited in the reviewed public material. Original visual by SultanByte.
ADNOC says the system was developed in the UAE and runs inside its cloud environment. SLB describes that environment as a sovereign cloud, with operational data and workflows kept in-country under UAE jurisdiction.
That is more specific than a generic data residency promise, but it does not answer every sovereignty question. Buyers should still ask who administers the platform, which support teams can access production data, where encryption keys are held, how software updates are approved, and whether operations can continue if the vendor connection fails. SultanByte's guide to choosing a GCC cloud region explains why physical location is only one part of control.
Industrial systems also need a clear boundary between recommendation and action. An alert that asks an engineer to inspect a pressure trend has a different risk profile from software that changes drilling parameters. The public releases do not say where ADNOC has drawn that line. Nor do they describe rollback procedures, model monitoring, or the evidence required before a recommendation enters a standard operating procedure.
These are not reasons to keep AI away from operations. They are the work required to put it there safely.
The RTOC is part of a broader programme rather than a standalone dashboard. Rigzone reported that ADNOC and SLB previously deployed an AI-powered production optimisation system across eight oil and gas fields, with a stated plan to extend it across all onshore and offshore fields by 2027.
ADNOC Drilling is also adding automation at the equipment layer. In June, the company introduced the AD-300, the first of six AI-enabled walking island rigs under a $1.54 billion contract. The company said the first rig arrived nearly three months ahead of schedule.
Together, those projects show three layers of an industrial AI stack: automated equipment, a fleet-wide operating view, and production optimisation across fields. The value will depend on whether data definitions, identity controls and incident processes work across all three. A technically impressive model will not fix conflicting asset IDs or unclear ownership of an alert.
The deployment gives CTOs and operations leaders a useful checklist.
Start with the workflow, not the model. ADNOC's clearest claimed gains come from replacing multiple tools, shortening reports and directing engineers toward problems earlier. Much of the benefit may come from integration and better information flow, with AI improving prioritisation.
Measure decision quality as well as speed. Hours saved are easy to market. False positives, missed events, unnecessary interventions and operator trust decide whether the system remains useful after launch. Buyers should require those measures before expanding from a pilot to a fleet.
Keep the evidence trail. Every recommendation should retain the source data, model or rule version, confidence level, operator response and eventual outcome. That record supports safety reviews and tells teams whether the claimed savings survive real operating conditions.
Design for degraded operation. Rigs cannot wait for a cloud service to recover. Local procedures, manual fallbacks and tested recovery paths matter as much as the central dashboard.
ADNOC's RTOC puts AI inside a large operational fleet. Its next proof point will be evidence that the reported gains persist over time, across different rigs, without increasing safety risk or concentrating too much control in one system.
Cover visual: Original technical illustration by SultanByte.
