AI: Energy seeing early AI efficiences. RTZ #556

...a vertical that shows horizontal implications for AI in businesses

The Bigger Picture, December 1, 2024

Investors have an eagle eye out on how long before we see returns in this AI Tech Wave, on the billions and trillions being expended on AI Data Center capex. WIth Nvidia being head of the line in the AI Gold Rush. The eagle eye is of course focused both on the consumer and enterprise markets. Across the board, horizontally, and into verticals like Healthcare, Education, Robotics, and many others. The Energy sector in particular is showing us the unexpected gains from AI even in these early days. And not just as a Power input for AI data centers. That is the Bigger Picture I’d like to focus on this Sunday.

Barron’s lays it out well in “AI is Changing Oil Country—and Pumping Up Profits”:

“The machinery in the vast Permian Basin is getting a lot more productive. And notably less noisy.”

“Over the past 15 years, oil drillers turned a stretch of desert in Texas and New Mexico called the Permian Basin into the most important oil basin in the world by re-engineering pipes and applying pressure and chemistry. Now they’re tapping a new technology to keep the crude flowing for decades more: artificial intelligence.”

“AI can feel like an unnecessary gadget in some industries—a chatbot to improve emails or spiff up spreadsheets. Not in energy. One expert thinks it can help companies extract so much more oil that it’s equivalent to adding the output of an entire Middle Eastern nation. “It’s like getting a Kuwait on-line,” said Rakesh Jaggi, who leads the digital business at SLB, the world’s largest oil-services company.”

The impact of AI is almost anti-climactic at first, since most expect world-changing impact at first blush. But the impact creeps up in pragmatic, organic, and bottom-up ways:

“Oil companies are on a relentless quest to produce more oil at lower costs, and AI is boosting that effort. Their success has already been remarkable. Over the past decade, the U.S. pumped out 60% more oil a day with 40% fewer workers. The industry’s annual productivity gains in that stretch outpaced even those of online retailers, and are roughly seven times as large as those of the average U.S. business. By extracting more oil while reducing capital expenses and manpower, they’re lowering the costs at which they can drill profitably. In the Permian, the “break-even” price for oil producers has fallen to $40 a barrel from over $90 in 2012, according to S&P Global Commodity Insights. AI should take that number even lower, boosting oil company margins and cash flow.”

“For the top Permian producers— Exxon Mobil, Chevron, Diamondback Energy, EOG and Occidental Petroleum —all of the extra cash they’re generating through efficiency gains should keep their dividends secure and growing, even during oil price slumps. Some of those stocks now yield over 4%.”

And data centers come into play even in this century old vertical:

“AI and oil might seem like a strange pairing. The sterile data centers that power AI programs have little in common with the grimy pipes and valves that pump crude. But oil drilling is a digital enterprise as much as a mechanical one today. Fields are seeded with sensors gathering reams of information about pressure, heat, radiation, and rock lithology. The action happening at the wells, tanks, and compressors is displayed on a screen hundreds of miles away—and it’s often being controlled from there, too.”

The industry does a whole lot more with less:

“At Chevron’s remote operations center in Midland, Texas, a couple dozen workers control thousands of pieces of equipment spread throughout the Permian. They each sit at workstations with 15 or so computer screens covered in blinking lines of data, interspersed with images of the fields captured by drones or still cameras. The technicians can remotely adjust the equipment themselves or direct people in the field to do it. Some equipment is even controlled autonomously by computers, using software that can detect pressure changes and adjust valves in real time. SLB says it has begun to deploy automated drilling operations, as well, making one of the most complex and stress-inducing portions of the process akin to riding in a robo-taxi. “You can sit back—not with a glass of Champagne, but maybe with a glass of juice,” Jaggi said.”

And AI is accelerating in use across many facets of oil and gas production:

“Tech advances ramped up in the past year or so. Chevron has been experimenting with artificial intelligence for years, but the company sped up its efforts last June as it became clear that the use cases were growing. “We began to see the future is essentially now,” said Steve Bowman, who leads a team of about 20 people focused on employing AI, particularly in the Permian region. Chevron built customized AI software using tools from Microsoft, including its Azure cloud platform.”

“Exxon has similarly begun to embrace AI after some initial caution. “We don’t like jumping on bandwagons,” said Exxon Mobil CEO Darren Woods on an earnings call last month. But he’s increasingly convinced that “AI is part of the equation” for Exxon’s future growth.”

Also, the scope is wide and deep:

“The Permian, which covers 75,000 square miles and accounts for nearly half of U.S. oil production, offers an excellent testing ground for AI. Drilling there is complicated, and better technology can make a noticeable difference. Unlike conventional wells where the oil is held in vast reservoirs, Permian oil is stuck in shale rock formations and has to be blasted out using water and sand. The hydraulic fracturing, or fracking, process involves drilling more than a mile down and several miles across through a wide range of rock types. Pressure in the wells declines quickly, causing production to fall as much as 70% in the first year.”

“Knowing where to frack can be an enormous advantage. The AI tools that Chevron has developed can predict which sections of the subsurface will yield the most oil, and which rock types will prove tricky to drill. Historically, oil companies have pulled out only about 10% of the recoverable oil in the sections of the Permian where they drill, a fraction of the recovery rate for conventional wells. Getting above that 10% level is considered the “holy grail” of Permian drilling, says Dan Pickering, founder and chief investment officer at investment and advisory firm Pickering Energy Partners. Bowman thinks the holy grail is now in sight. “It’s a good opportunity for AI,” he said.”

All of this is changing how Energy works contrary to popular perceptions:

“Modern advances in oil production don’t line up with “drill, baby, drill,” the rallying cry of President-elect Donald Trump. Fifteen years ago, the early Permian shale producers used that kind of brute-force method. They raced to lease and drill land, burning through almost $400 billion in cash, scaring off investors, and sinking the stocks. It’s one reason that oil stocks are still unloved today, making up less than 4% of the total S&P 500 index market cap.”

And these companies don’t have to spend tens and hundreds of billions on AI today for returns tomorrow.

“It isn’t easy to pinpoint how much of the advancements that oil companies have made is from AI. Oil-and-gas companies are on track to spend $3.1 billion on AI this year—less than 5% of their capital expenditures—but annual AI spending could rise 80% in the next five years, according to market research firm Mordor Intelligence. SLB is seeing a surge in interest in these tech tools, and expects to surpass $3 billion in revenue from its digital division next year, up from $1.5 billion in 2021, Jaggi said. Its digital services include helping producers adopt AI.”

Sensors, an important new input for AI going forward, are playing a big role in AI driven exploration and drilling, as well as horizontally across verticals:

“Some of the most impressive gains are happening in offshore oil drilling. To find oil underwater, producers use seismic imaging, which involves shooting compressed air into the water, and using sensors placed on the sea floor or tugged behind boats to evaluate the acoustic waves that are sent back. Evaluating the images is a time-consuming and inexact science. Jaggi of SLB said that the time lag between doing seismic exploration and determining where to drill used to be about a year and half. AI has changed that dramatically.”

The whole piece is worth reading in full for additional nuance on the impact of AI in this vertical already. But the Bigger Picture here is that AI impact is already important bottoms up in verticals large and small.

Energy in this case is just a preview on the impact to come across most verticals sooner than later, regardless of ebbs and flows in AI sentiment.

It’s the ‘Enterprise Surplus’ to be expected with AI generated efficiencies, just like ‘Consumer Surplus’ has driven tech efficiencies for decades. AI is likely to accelerate both types of surpluses, as this AI Tech Wave progresses. Stay tuned.

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)