Data Centers in the Permian Basin Part 2 – It’s More Than Just Trapped Gas
Implications for LandBridge, TPL, Vantage Data Centers, & DigitalBridge
In August of this year, I wrote the article Data Centers in the Permian Basin (https://310value.substack.com/p/data-centers-in-the-permian-basin). Subsequently, there have been some events that make the thesis that data centers will be in the Permian Basin more likely to come true. Additionally, other factors that would help push hyperscalers to locate data centers have become apparent. Let’s start with some recent events.
The first event was the announced restart of Three Mile Island nuclear power plant. Shortly after the restart announcement, a regulatory review was launched. The review could cover areas that are site specific, such as how the restart of Three Mile Island’s Unit 2 would coincide with the decommissioning of Unit 1. It could also cover areas such as the safety of nuclear energy in general or ramifications of electricity crossing stateliness. While I believe that nuclear energy will be a key solution to our future energy needs, I have two conclusions regarding the announcement and regulatory review. First, nuclear will take time to commission; and second, from a regulatory perspective, it is best if electricity does not cross state lines.
The second interesting event was LandBridge’s letter of intent (“LOI”) going definitive. The definitive LOI is for a two-year site selection and pre-development period. The counter party is a partnership between LandBridge’s sponsor, Five Point Energy (“FPE”), and “an affiliate of Silver Lake”. This announcement has multiple ramifications, so let’s unpack it. First, let’s look at the partnership. We can discount FPE given their relationship with LandBridge, but the affiliate of Silverlake is interesting. This is a third party writing a check. The only data center business that I am aware that Silver Lake is affiliated with is Vantage Data Centers. Vantage is partially owned by DigitalBridge. Vantage Data Centers is a respected operator, so if it does happen to be involved, that lends credibility to the Permian data center thesis. Additionally, I don’t think Vantage is in the habit of spending millions of dollars for a pre-development lease unless there is a real opportunity. While this is still in the proof of concept phase given that the LOI is basically a two-year call option on development of a data center, given the sophistication of the buyers of that option, it is worth digging into why it makes sense for data centers to go to the Permian Basin – Taylor Sheridan pay attention; I expect a rant on this by Billy Bob Thornton’s character in Landman Season 2.
Why AI Data Center Make Sense in the Permian Basin
Energy
Trapped natural gas provides a cheap fuel source of electricity generation.
Ability to create a microgrid and not create a burden to an aged transmission system.
Furthermore, these microgrids will allow data centers access to power without competing with individual electricity consumers; rising energy prices due to data centers is something that will invite regulatory scrutiny.
Access to renewables. Hyperscalers have net zero goals. The large energy consumption of AI data center could derail the net zero goals of hyperscalers. The placement of data centers in the Permian Basin alleviates this by having ample land available for wind and solar energy production.
Water
AI data centers consume massiveamounts of water for cooling. The below article (2) from the World Economic Forum states that a 1 MW data center consumes 25.5 million liters of water annually; this is equivalent to the water needs of 300,000 people.
Not only does the Permian have brackish water available, but it also has produced water from oil and gas production that can be cleaned and recycled. The use of recycled produced water will also help hyperscalers in their ESG goals. Note, given the cost to clean produced water, I only see its use as a way to adhere to self-imposed ESG goals, not as a cheap source of water.
Speed of Development
Locating data centers in the Permian Basin can de-risk the time to development.
Permitting a data center in the US can take 24-36 months, in the Permian this can be shortened by half.
It’s worth noting that there is a data center project in California that is 4.5 years into the permitting process and a shovel has yet to get dirty. DigitalBridge’s Marc Ganzi echoed this sentiment in a recent podcast (3). As a side note, Ganzi also raises the issue of water for cooling several times on the podcast.
Energy and water, not to mention no nimby issues, are two reasons why locating AI data centers in the Permian Basin makes economic sense. That said, speed to development might be a more important factor. Hyperscalers seem to be in an AI arms race and shortening the time needed for permitting and development helps keep hyperscalers ahead of the curve. Time being the most important constraint also helps answer the question of why hyperscalers will not do this themselves. While permitting time is cut down in Permian, it doesn’t mean that there won’t be requisite easements and entitlements required to put up a data center campus. Furthermore, a campus will need consistent access to water. Outsourcing these functions seem to be the way to go and is likely why hints of the first major Permian project are coming from LandBridge.
Implications for Potential Players
LandBridge: LandBridge is the first to get mentioned, as they are the first to disclose a definitive LOI for a project. If the data center thesis plays out, it can mean a boost to their earnings stream. They have several sites available for data center campuses, including the recent Wolf Bone acquisition, and I’ve seen estimates of annual revenue in the range of $30M to $40M per megawatt of data centers.
Texas Pacific Land Corporation (“TPL”): TPL is the 800-pound gorilla when it comes to Permian surface land, that said, they need to put together ~2,000 contiguous acres per data center campus. On the Q3 2024 earnings call, the TPL CEO mentioned the opportunity, but they will need to start putting together surface acreage to make this happen given their checkerboard surface ownership.
Other Permian Oil and Gas Operators: I can think of at least two oil and gas companies with land mentioning the opportunity on Q3 2024 earnings calls. While these operators may have land, I see them as likely partnering with a data center expert to realize this opportunity. One reason they would partner is that they do not have the in-house expertise to execute on this new opportunity. Another reason would be a partnership would likely produce cash for them to use in their primary business – drilling for oil.
Vantage (DigitalBridge): It would be a feather in the cap for Vantage and those associated with it to get a data center developed in the Permian. That said, given the size of these businesses, I see it as another place for them to deploy ample amounts of investment capital. One could easily see the size of investments in the Permian Basin grow to $50B to $100B per campus. Scala, a DigitalBridge backed company, announced a $500M investment in a campus in Brazil, with aspirations to grow it to $90B in total investments (4).
Notes:
2. https://www.weforum.org/stories/2024/11/circular-water-solutions-sustainable-data-centres/
3.
4. https://www.datacenterfrontier.com/site-selection/article/55140968/brazils-scala-data-centers-secures-latam-dominance-plans-to-develop-ai-city
Disclaimer: For entertainment purposes only. Not financial advice. Not a solicitation to buy or sell any security. Do your own due diligence.
How do you think Microsoft’s new design for AI data centers to run without water needed for cooling will play into these companies long term prospects of having data centers built on their land?
The giant land parcels that LandBridge has acquired have a very small number of residents. Can we be certain the engineers required to run data centers actually want to live in such remote areas?