Imagine a world where the rapid advancement of artificial intelligence (AI) hits an unexpected roadblock—not due to a lack of innovation, but because of something as mundane as underground salt caverns. Yes, you read that right. These little-known geological formations could be the key to sustaining the AI boom, yet their scarcity is now threatening to derail the race against global competitors like China. But here’s where it gets controversial: while the world rushes to build pipelines and power plants, the critical infrastructure for storing natural gas—the lifeblood of AI’s energy demands—has been largely overlooked for over a decade. And this is the part most people miss: without reliable gas storage, even the most advanced data centers could grind to a halt during supply disruptions.
The issue is simple yet profound. Natural gas, a primary energy source for AI-powered data centers, relies heavily on pipelines and storage facilities. However, pipelines are vulnerable to weather events, landslides, and corrosion, leaving facilities at risk of power outages even when gas is available elsewhere. Edmund Knolle, president of Gulf Coast Midstream Partners, warns, ‘Everybody in the data center world seems to be in a great big hurry, but they haven’t fully considered the risks on the gas side.’ His company is developing a major salt cavern storage project near Houston, slated for completion by 2030, but it’s just one piece of a much larger puzzle.
Energy costs are already soaring due to higher gas demand, and analysts predict that the lack of storage will exacerbate volatility and drive utility bills even higher. Knolle bluntly states, ‘We’re going to be way short of storage. Once people realize this, it’ll take years to catch up.’ The situation is dire, yet progress is slow. Enbridge, North America’s largest pipeline and energy storage company, is expanding its salt caverns in Texas and Louisiana, but even they admit the challenge is growing. ‘Our pipes are full, and demand is unprecedented,’ says Caitlin Tessin, Enbridge’s vice president for gas transmission. ‘There’s real concern about storage capacity.’
Here’s the kicker: Natural gas pipelines and storage are poised to become the ‘backbone’ of AI’s digital infrastructure, even complementing renewable energy to ensure 24/7 reliability. But with only a handful of storage projects underway—many lacking financing or years from completion—the gap between demand and supply is widening. Energy analyst Jack Weixel estimates that twice the currently planned storage capacity is needed to meet the demands of data centers, electrification, and gas exports. ‘The number one rule for utilities is don’t freeze grandma,’ he quips, emphasizing the need for a stable grid.
Salt caverns, created by drilling into natural salt domes and injecting water, are the gold standard for gas storage due to their durability and high-pressure capacity. However, the process is time-consuming, often taking four years. A cheaper alternative involves using depleted gas reservoirs, but these are less reliable and cannot handle frequent injections or withdrawals. Worse, weather disruptions—especially hurricanes targeting the Gulf Coast—could strand gas supplies, as storage is already near capacity nationwide. ‘It’s operational chaos,’ Weixel warns.
Efforts are underway, but they’re a drop in the ocean. Knolle’s Freeport Energy Storage Hub (FRESH) project aims to add 26 billion cubic feet of gas capacity by 2026, while Enbridge is expanding facilities in Texas and Louisiana to add 23 billion cubic feet by 2033. Trinity Gas Storage’s East Texas facility, one of the few recently completed projects, offers 24 billion cubic feet and plans to expand by 13 billion feet by 2026. Yet, these projects are just the tip of the iceberg.
Here’s the controversial question: Is the industry moving fast enough? Trinity CEO Jim Goetz believes capitalism will find a solution, but he admits, ‘We’re caught behind the eight-ball.’ The last gas storage boom ended around 2010, and since then, progress has stagnated. Now, with AI’s insatiable energy demands, the race is on to build ‘gas storage super-cycle 2.0.’ But will it be enough? And at what cost to consumers and the environment?
This isn’t just a technical challenge—it’s a call to action. As Goetz puts it, ‘It’s a problem that must get solved.’ But how? And who will lead the charge? The answers may determine the future of AI—and our energy security. What do you think? Is the industry doing enough, or are we headed for an energy crisis? Let’s debate in the comments.
