The frenzied build-out of artificial intelligence (AI) data centers in the U.S. is going to require gas-fired electricity everywhere, including in California and the Northeast, where existing delivery infrastructure is already running at near full capacity.
Unfortunately, policy choices create choke points blocking development of new delivery capacity to these regions. A “choke point” in this context is a place where the physical capacity of a pipeline system is too small to meet demand. Choke points limit the ability to move natural gas to businesses and families, which increases prices and reduces reliability.
In other words, these policy-driven supply constraints make electricity, heating, and cooling more expensive while simultaneously making it more likely for customers to experience brownouts or blackouts.
A “brownout” is a temporary reduction in voltage — demand that exceeds an electrical grid’s capacity may cause lights to dim and electronics to malfunction. A “blackout” is a total outage that stops all electrical transmission.

Looking at our nation’s natural gas distribution system, it might seem that we have the ability to reliably deliver gas from producing regions to customers. In most cases, that’s true. The exceptions are places such as California and the Northeast, where supply is limited by these choke points.
With few exceptions, economic growth and energy consumption are linked. The AI boom is forecasted to dramatically increase energy demand and economic growth, which will magnify the impact of existing choke points.

A few facts:
- The International Energy Agency (IEA) reports that data centers in the U.S. comprised about 2% of electricity demand in 2018. Data center demand has more than doubled to about 4.5% and is expected to double again to 9% in 2030. Existing non-AI demand is not expected to decrease.
- Data centers need cost-effective, reliable electricity that can be provided within the needed timeframe only by natural gas-fueled systems.
- Natural gas power generation is by far the largest source of electricity in the U.S. and is increasing — even without the AI data center build-out.

The point: electricity demand is going up, which means the need for natural gas-fired generation is also going up. Those facts, and the policy-driven choke points in California and the Northeast, are a recipe for skyrocketing energy prices and decreasing reliability in those regions.
The California Problem
- High utilization of natural gas transmission pipelines into and within California places a serious limitation on new gas-fired generation built in support of AI data centers.
- “In Southern California, Citygate and Redwood path constraints north to south from PG&E will oftentimes leave Southern California in a lurch for gas when demand is high, particularly in the summer,” wrote Jack Weixel, Senior Director, Energy Analysis at East Daley Analytics.
The Northeast Problem
- According to the U.S. Energy Information Administration, natural gas production in the Appalachian Basin is limited by pipeline export capacity into New York and New England. In other words, Pennsylvania, West Virginia and Ohio could produce more natural gas to meet their neighboring states’ energy needs if there were adequate pipeline capacity.
- According to Utilitydive.com, “A recent reliability report from the North American Electric Reliability Corporation (NERC) and Northeast Power Coordinating Council (NPCC) found that natural gas infrastructure in New York and New England is fully utilized during extreme winter weather and that lack of spare pipeline capacity poses severe threats to reliability.”
- “Every winter, the capacity into New England at the Algonquin Hub gets maxed out when it gets cold. There simply is no room to effectively move more gas west to east across the Hudson River, and Algonquin prices blow out 5-10x normal price,” wrote Weixel.
- To meet their power and heating needs, New York and New England are forced to burn fuel oil (which produces more greenhouse gas emissions than natural gas) or import liquefied natural gas, which is 2-5 times more expensive than natural gas depending on market conditions.
There are other choke points impacting natural gas distribution throughout America, but the ones in California and the Northeast are the biggest problems. Given that gross domestic product (GDP) of these two regions is approximately 27% of U.S. GDP, and substantial AI-driven economic growth in these regions is desired, it’s in the national interest to eliminate choke points to better serve these regions.
Luckily, these are solvable issues. Midstream companies are ready to collaborate with federal, state, and local governments to design balanced energy policies that reflect the need for many ways to make energy, including natural gas and renewables like wind and solar.
We can build the infrastructure to meet our needs. If we choose not to, California, New England, and the U.S. as a whole will not fully benefit from the economic growth offered by AI.
Thomas Kalb is Director of the Coastal Bend Midstream Program at Texas A&M University-Corpus Christi. He is a regular contributor to the Let’s Clear the Air blog, where he writes about issues, policies, and considerations impacting how we produce and use energy.
Share This