BusinessJul 14, 2026
AI's Next Gold Mine Lies in the Oil Patch
Physical Scaling
Rhea Lobo
Discover
The companies racing to build AI infrastructure aren't just tech giants anymore. Energy firms with deep experience in rapid field deployment are moving into data center construction and power supply.
That shift is now attracting billions in private capital and reshaping who benefits from the AI boom.
SLB, the world's largest oilfield services provider, partnered with Liberty Energy to supply modular prefabricated components and natural gas-fired power to data centers globally.
SLB will handle the physical modules and Liberty will supply the power generation. Liberty CEO Ron Gusek pointed directly to the skills transfer from their existing business.
"We've demonstrated an incredible nimbleness and responsiveness that's allowed the oil and gas industry to evolve at the pace it has. We're bringing that same DNA to the data center space now."
Ron Gusek, Liberty Energy
For Liberty, the move is also a hedge. Providing data center services creates a revenue stream that isn't tied to oil prices or drilling activity, which can swing sharply.
SLB is already a design partner for modular AI data centers built on Nvidia technology. The two companies are working together on a platform called AI Factory for Energy, designed to help oil and gas producers apply AI to operational data.
At the heart of this shift is hyperscalers' race to build AI infrastructure. They're committing enormous sums to new data centers and need power and buildings as quickly as possible.
Meta announced its Hyperion data center supercluster in Richland Parish, Louisiana will cost more than $50B, up from the $27B figure disclosed recently. The project started at a $10B estimate and has grown sharply since then.
That pattern of expanding commitments is not unique to Meta. Hyperscale cloud operators including Amazon, Microsoft, and Alphabet have collectively announced more than $200B in combined data center investment commitments.
The global data center rack market alone is projected to grow from roughly $7.5B in 2026 to $13.5B by 2033, a compound annual growth rate of 8.8%.
AI workloads are the main driver of that growth. Modern AI GPU clusters consume between 30 kW and 100+ kW per rack, far above what traditional enterprise servers require. That pushes demand toward higher-density, reinforced cabinet racks with liquid cooling integration.
With construction costs rising and demand unrelenting, many data center operators are selling equity stakes to fund their next expansion phase.
US data center builders are working with bankers to offload majority stakes worth tens of billions of dollars this summer.
Firms including Netrality Data Centers, DataBank, Edged, and EdgeCore Digital Infrastructure are among those in active sale processes.
In 2025, data center mergers and acquisitions hit roughly $50B, more than double the prior year. Blackstone recently took a ~49% stake in Rowan Digital Infrastructure.
Not every deal will close. Ravi Purohit, co-head of infrastructure at Paul Weiss, noted that while investor appetite is strong, the buyer pool for multibillion-dollar checks is thin.
Rising construction costs, power access constraints, and local community opposition are adding risk to deal valuations. Buyers are increasingly tying payments to milestone completions rather than paying upfront.
The physical buildout of AI is accelerating and the money flows are broadening well beyond software and chip stocks. Energy service companies, infrastructure funds, and rack manufacturers are all moving toward the center of that trade.
