When the rig starts thinking: The rise of autonomous drilling

Posted: September 29, 2026

Thousands of feet below the surface of the sea, a drill bit is cutting through rock that no human has ever seen. Conditions can change as it descends: pressure fluctuates, rock layers vary and the bore slips from its planned trajectory. All the while, streams of data travel back towards the surface. But increasingly, it isn’t only a human driller interpreting that information and deciding what to do next.

In a recent deepwater drilling campaign, an artificial-intelligence-driven autonomous system was connected to the drillship’s rig-control system and downhole tools. Three automation systems exchanged information onboard, enabling them to coordinate in a closed loop.

The project, involving Eni, SLB and Saipem, was presented as the first global application of autonomous drilling in deep water. But it’s only one of a growing number of projects suggesting that drilling automation is entering a new phase.

Automated drilling itself is not new. But oil companies are increasingly connecting previously separate systems, from rig controls and directional drilling to subsurface interpretation and pressure management. As the industry’s automated tools start talking to one another, where does drilling automation end—and genuine autonomy begin?

From drilling automation to autonomous drilling

At its simplest, automation allows a machine to execute predefined tasks or workflows. Autonomy goes further: systems can use incoming information to assess conditions, determine a response, act within defined limits and use new data to evaluate the result. The ambition is to create a closed loop: sense, interpret, decide, act and measure again.

Better sensing is an important part of that equation. Distributed fiber-optic sensing can provide continuous information along wells and other oil and gas infrastructure, helping overcome some of the limitations of conventional point sensors. Data analysis can turn those measurements into information that operators—and increasingly automated systems—can act upon.

The challenge is to connect that intelligence with the machinery doing the work. Eni’s deepwater project integrated its AI-driven autonomous system with two other automation systems onboard, while also connecting it to the rig controls and downhole tools. Together, the systems could exchange information and coordinate drilling activities.

Eni has since pushed the integration further. In July, the company and Halliburton announced the deployment of closed-loop rig automation on a deepwater exploration well offshore Indonesia. The project linked equipment at the surface with automated well placement, downhole hydraulics and pressure-control systems.  

Halliburton’s orchestration software coordinated drilling and tripping while integrating pressure management into the same automated process. “When we integrate rig automation with [managed-pressure drilling] controls, we close a critical gap and give operators better control, consistency, and performance in complex wells,” says Jim Collins, vice president of Sperry Drilling at Halliburton. The company says the approach improved efficiency by more than 15% while maintaining well control within a narrow operating margin.

How oil and gas companies are optimizing drilling in real time

A similar transition can be seen at bp’s Atlantis field in the Gulf of Mexico, where, last October, the company automated managed-pressure drilling, which precisely controls pressure in a well. Real-time data allowed the system to optimize drilling “on the go”, rather than relying solely on parameters determined in advance.

The first run delivered what former bp senior executive Ann Davies described as stable and consistent drilling, and gave the company confidence to scale the technology. bp has also been using algorithms for early kick detection, moving towards anticipating changes in well conditions rather than simply responding to them.

The direction of travel is clear, according to Davies. “Eventually, the destination is full closed-loop automation,” she told the Journal of Petroleum Technology.  

But closing the loop around how a well is drilled is only part of the equation. A more difficult challenge is allowing automated systems to influence where it goes.  

Offshore Guyana, ExxonMobil and Halliburton have taken a step in that direction. Working with several other partners, the companies announced in March what they described as the industry’s first fully automated geological well placement with complete rig automation.

The project brought geological interpretation, well-placement decisions, hydraulics and automated rig controls into the same system. Geological data could therefore inform both where the well was steered and how the rig executed the drilling in real time.

Another ExxonMobil project off Guyana, this time involving SLB, shows the scale at which autonomous directional drilling can already be deployed. SLB says its technologies have autonomously navigated more than 48 km of complex three-dimensional well trajectories, with the operations overseen remotely from an onshore center.

More than 93% of operations were executed autonomously, according to SLB, with the required downlinks—the commands and signals travelling between the surface and the well-bore—decreasing significantly. The result is a much larger closed loop, extending from information about the well and its trajectory to the machinery executing the actual drilling.

How autonomous drilling is scaling across multiple wells and rigs

While these technological firsts demonstrate what’s possible, a bigger test will be whether autonomous drilling can deliver consistently across different wells, rigs and fields.

Evidence published by Halliburton from the Norwegian Continental Shelf suggests that autonomous directional drilling is beginning to build that kind of operating record. Halliburton compared three years’ worth of experience with conventional operations across four rigs. According to the company, the automated sections achieved an average rate of penetration of 36.71 meters per hour, compared with 29.75 meters per hour for conventional drilling—an improvement of 23.4%.

Halliburton also reports improvements beyond drilling speed. Its analysis found reductions of up to 15% in casing and liner running times, while the autonomous sections showed reduced wellbore tortuosity—the unwanted twists and deviations in a well’s path—and around four times more meters drilled between failures.

AI connects drilling operations across entire rig fleets

If the Norwegian results show autonomy working across multiple wells and rigs, the next step is connecting operations at fleet scale. In August, ADNOC and SLB announced they had deployed an AI-enabled real-time operations center across more than 120 onshore and offshore drilling rigs.

The platform brings real-time operational data together with advanced analytics and AI-driven insights, helping drilling teams spot potential problems sooner and make faster, better-coordinated decisions in response.

The platform does not make those rigs autonomous; it aggregates, displays and analyzes data. But ADNOC and SLB position it as a foundation for more autonomous operations, connecting data and AI-enabled workflows across a large rig fleet.

Already, ADNOC says the human effort involved has fallen by 30–40%, with individual engineers now able to oversee two to three times more rigs. Some analytical work that once took a day can now be done in minutes.

As these recent examples show, the autonomous rig is unlikely to arrive through a single breakthrough. Instead, it’s emerging as individual loops connect. Real drilling autonomy will come from connecting these capabilities—allowing the different systems that sense, interpret, decide and act to work together as an increasingly integrated whole.