Data Centers in the Permian Basin

Can Texas Pacific Land Corp. and LandBridge Capitalize on Trapped Natural Gas?

The top three discussion points when discussing data centers these days are power, power, and power.  Earlier this year at Citi’s Global Property CEO Conference, DigitalBridge’s CIO, Ben Jenkins, noted that the public cloud has roughly 300 gigawatts of capacity, which took ten years to deploy.  Ben and his colleagues at DigitalBridge believe AI will reach a similar size in just two to three years.  It is not just data center operators talking about this.  Kinder Morgan Executive Chairman, Richard Kinder, on Kinder’s Q1 ’24 earnings call noted that data centers currently use 2.5% of US electricity, but power usage from AI alone is expected to be about 15% of total electricity demand by 2030.  Texas Pacific Land Corp. (TPL) and LandBridge Company LLC (LandBridge), both of which own massive amounts of surface acreage in the Permian Basin, are two companies poised to benefit from this.  In fact, LandBridge’s chairman, David Capobianco, discussed this on a CNBC interview he gave the day of LandBridge’s recent IPO (see below).  Additionally, LandBridge announced on their first earnings call that during the second quarter, they signed a non-binding letter of intent (LOI) for the development of a data center on their land in Reeves County.  The earnings release noted that the LOI contemplates an $8 million fee payable to LandBridge in January 2025 upon the execution of long-term ground lease.  This raises the following questions: What makes the Permian Basin suited for data centers; what kind of data centers would come to the Permian Basin; who would operate the data centers in the Permian Basin; and what are the economic implications for Permian Basin land companies such as TPL and LandBridge?

David Capaobianco CNBC Interview

What Makes the Permian Basin Well Suited for Data Centers?

As noted in the LandBridge IPO video, the Permian Basin has some of the cheapest natural gas in the country, if not the world.  Additionally, there is plenty of land available for wind and solar farms, as well as for locating data centers.  Furthermore, processed produced water from oil wells can be used as a cooling source for data centers. 

What Kind of Data Centers Would Come to the Permian Basin?

Data centers need energy.  As it relates to the Permian Basin, there are a few options.  First, you can build power generation facilities in the Permian Basin and transport power outside of the area for use in a data center.  However, given the precarious state of Texas’ electrical grid (specifically ERCOT, which serves 75% of the state), and its vulnerability to grid failures, this isn’t a likely solution in the short to intermediate term.  Second, you can build a natural gas pipeline to transport natural gas out of the Permian Basin to places where it is needed.  Building a natural gas pipeline is a long, costly, difficult process that takes years to complete.  It is highly regulated, and the number of regulations multiply when state lines are crossed.  Natural gas pipelines are being added to the area, but it is much easier, cheaper, and quicker to transport data than it is to transport natural gas.  This leads us to the third option, locate a data center in the Permian Basin.  Such a data center can be located on its own microgrid and be “behind the meter” so to speak, as it can be collocated with its own power generating equipment, whether it be a solar or wind farm, and or a natural gas power plant.

Back to the type of data centers that would work there.  Retail colocation data centers likely have latency requirements that would prevent a company from locating a data center in West Texas and then transporting data to a metro area.  Furthermore, the operators of these data centers (Equinix, Digital Realty, etc.) like their data centers to have a diversity of uses and customers, which you would not find if you located a data center in West Texas.  Hyperscalers, who take whole floors of data centers if not the entire data center, likely have similar latency requirements of the retail colocation data centers for some of their applications, so the Permian Basin would not be ideal for all hyperscaler applications.  The one exception would be the large language models of generative AI.  Thus far, these applications tend to not be latency sensitive.  The Permian Basin could be a home for AI data centers. Hyperscalers can place their own data centers in the Permian to satisfy some of their AI processing needs or they can partner with third parties. It sounds like LandBridge’s financial sponsor is keen to be finance data centers placed on LandBridge’s surface land. Additionally, digital infrastructure investor DigitalBridge already has experience executing similar strategies with their portfolio companies Switch and Scala, with projects in Reno, Spain, and Brazil. I could see them doing something in the Permian with either portfolio company Switch or Vantage.

With three types of energy available in the Permian Basin (wind, solar, and natural gas), it is likely that a data center that is collocated with electricity generation will have more electricity than it needs.  This makes the location of TPL’s and LandBridge’s land interesting.  Excess electricity produced on their borderline land should be able to be sold to the ERCOT and WECC grids in Texas and Xcel Energy’s grid in New Mexico. 

LandBridge’s Announcement

As noted, LandBridge announced a non-binding LOI on their first earnings call for a data center.  What type of data center is it?  We will learn more if the LOI becomes binding.  Until then, let’s speculate a bit.  The best news for LandBridge, and its shareholders, would be if this LOI is with a hyperscaler that wants to locate an AI data center on LandBridge’s property.  This would be a first step in proving out part of the investment thesis in that company (note, there is plenty to be excited in LandBridge setting aside the data center part, but the data center part of the thesis certainly has some sizzle).  Aside from an AI data center, the other option that makes sense would be a fringe use such as a bitcoin miner.  Bitcoin mining needs cheap energy, so co-locating a bitcoin miner in the Permian makes some sense, assuming the operation can amortize the cost of the power generation facility.

Financial Implications to Permian Land Companies

There is already the LOI for LandBridge in the amount of $8 million if the LOI becomes binding.  Presumably, there will be ongoing ground lease or royalty payments made.  Additionally, LandBridge, or their sponsor, could take an ownership position in any electricity generation facilities, which along with water for cooling, could also provide ongoing revenue sources.

Despite LandBridge’s rapid share price rise since its June IPO (up 143% as of the morning of 8/23/24) and TPL’s continued strong performance, I would argue that very little of that stock price performance has to with both companies’ cheap call option on the data center opportunity.  That said, with the potential eight-figure annual checks (with no accompanying cogs), the data center opportunity could provide compelling value drivers to both stocks in the near future.  It would be best for both companies if the data center partners were hyperscalers that were building AI focused data centers, as opposed to fringe uses such as crypto miners.  That said, with large language models not currently requiring low latency, and the cheap power in the Permian, it makes sense for hyperscalers to start building data centers there.     

Disclaimer: For entertainment purposes only. Not an offer or a solicitation to buy or sell any security. Do your own due diligence.