One of the Best-Performing S&P Sectors Has an AI Story Nobody Is Telling

Listen to the audio version of this article (generated by AI).

Editor’s Note: If you read Joe Austin‘s piece yesterday, you know the argument: the real AI money isn’t where everyone’s looking.

Today he’s back to prove it again, this time in a different industry entirely.

Same thesis, new terrain — and the opportunity is just as overlooked. Joe and Marc Chaikin are laying out the full picture on Wednesday, June 24, when they’ll debut the first AI-powered tool Chaikin Analytics has ever built.

But charter access is limited, so reserve your spot while you can.

Now here’s Joe…

The search for oil and gas never stops. And it never looks the same.

On Alaska’s North Slope, rigs operate in some of the most punishing conditions on Earth. In winter, temperatures routinely drop well into the negatives. Around the winter solstice, daylight can last as little as two hours a day.

The enemy is the environment. Low light, brutal cold, and encroaching sea ice shut down operations for nearly half the year. When your drilling window is that short, every hour counts.

Meanwhile, thousands of miles south, drillships in the warm waters off Guyana operate over the Stabroek Block — a deepwater tract that ExxonMobil Corp.’s (XOM) CEO has called one of the biggest oil discoveries in nearly two decades. The water here can be more than 6,000 feet deep before you even reach the oil reservoir. 

Everything costs a fortune, and nothing stops. The day rate for drillships runs between $400,000 and $500,000 per day.

Then there’s the U.S. shale patch.

In the Permian Basin and the Marcellus Shale, the challenge isn’t weather or day rates. It’s doing more with less. From late 2022 through late last year, the active rig count in the lower 48 states dropped by about one-third.

But over that same period, Permian production jumped 18%. Appalachia production increased 10%. And last July, the lower 48 states set a new monthly production record for crude oil.

Fewer rigs. More oil. That’s efficiency — and AI is driving it.

These are three environments with very different problems. But the solution is always the same: better technology. And right now, that means AI.

AI Is Reshaping Oil and Gas Drilling In Real Time — and the Results Are Measurable

Across the oil and gas industry, AI is reshaping how wells get drilled.

The basic machinery has been around for decades: a derrick to support the drill string, a rotary system to spin the bit, a hoist to raise and lower equipment, and a circulation system to pump drilling fluid in and out of the hole. But what happens inside those systems has changed dramatically.

Sensors in the drill string now send live data up from the bottom of the hole while drilling is still underway. That gives engineers a real-time read on rock type, pressure, and well direction. Software tracks mud weight and chemistry in real time, catching pressure warning signs before fluids start flowing into the well uncontrolled.

Directional drilling lets crews bend the well path underground to reach targets thousands of feet away — making it possible to drill multiple wells from a single surface location. And as each section is drilled, it gets lined with steel casing and cemented in place. Evaluation tools verify the cement has set before the crew moves deeper.

For years, skilled operators and engineers managed all of this by reading data, making judgment calls, and adjusting on the fly. Now, AI is taking over that work. And the results are measurable.

The AI Revolution That’s Already Happening Where Nobody’s Looking

Surface systems no longer just follow preset rules. They learn from live well data, make decisions, and adjust drilling parameters faster and more consistently than any human can. In one 2024 drilling program, an AI-driven system drilled nearly 50% faster than a manual crew.

Downhole, AI now interprets data from drilling tools in real time and adjusts the well path automatically — keeping the bit in the most productive zone without waiting for a geologist to weigh in. At a well in Ecuador, an AI system made 25 course corrections along a single well section, each in seconds. That well became one of the best producers in the country.

On the fluids side, machine-learning models can flag signs of a pressure imbalance 10 to 12 minutes earlier than conventional monitoring tools. And cement evaluation models that once required a specialist to manually read complex acoustic logs now run automatically, faster, and with better accuracy.

This is what physical AI looks like. It’s not a chatbot or a software upgrade. It’s machines making real-time decisions in conditions where a human mistake costs millions of dollars — or worse.