"Could the Global AI Boom Trigger a 'Gas Crisis'? Expert Warns of Depletion by 2030 [Tech Talk]"

Gas Turbines: The Heart of AI Data Centers
Natural Gas Demand Set to Surge Soon
"A Choice Between AI and Consumers"

Editor's NoteThis series unpacks complex and often unfamiliar technologies—ranging from AI, semiconductors, telecommunications, to biotech—in simple and easy-to-understand language, showing how they are essential to our lives.

"By 2028, we could face a natural gas shortage, and by 2030, the United States' gas storage facilities might be depleted."

This warning was voiced by Matt Smith, an American energy resources analyst, during a recent appearance on the well-known podcast 'Invest Like the Best.' Having spent decades analyzing fossil fuel market trends, he says he has observed an unprecedented surge in natural gas demand in recent years. The driving force, he explains, is the global artificial intelligence (AI) data center boom.

Today, among the various types of power generation facilities, data centers are particularly compatible with gas turbines. As more data centers operate on natural gas, ordinary people may find themselves in a situation where gas for heating becomes scarce.

AI Boom Triggers Soaring U.S. Natural Gas Demand

Drilling facility in Texas, the largest shale oil production area in the United States. Photo by Yonhap News Agency

Based on Henry Hub, the world's largest natural gas and liquefied natural gas (LNG) trading market in the U.S., the price of natural gas remains around $2.70 per million BTUs, which is extremely low compared to average levels. However, Smith warns that natural gas demand has surged sharply in recent days, and by 2028, bottlenecks could spill over into higher prices.

The main culprit behind this surge in natural gas demand is AI-focused data centers—a point already highlighted in a report released in May by the International Energy Agency (IEA). The IEA noted that gas turbine generators are multiplying to supply electricity to AI data centers, and this could cause annual investments in fossil-fueled power generation in the U.S. to surpass China for the first time in decades. By 2028, the estimated power consumption of data centers scheduled to be built in the U.S. is projected to range between approximately 130 to 150 gigawatts (GW).

Currently, Industrial Gas Turbines Are the Heart of Data Centers

Industrial gas turbine from General Electric (GE) installed in an artificial intelligence (AI) data center. CNBC YouTube

Data centers, packed with computer chips and cooling equipment, are well-known as massive electricity consumers. Recently, even top executives at various AI companies have cited power supply—not chips—as the true bottleneck for AI development. Yet, out of all the power generation options, data centers pair particularly well with industrial gas turbines.

Why is that? For one thing, data centers must operate non-stop, 24 hours a day, 365 days a year. Renewable energy sources, whose output varies with weather, pose a risk in this context. In contrast, gas turbines deliver much higher output than other generation options and are easy to adjust for load. Most importantly, gas turbines can ramp up to full output quickly after startup, meaning their warm-up time is extremely short. If the power grid fails, backup gas turbines can be spun up instantly to restore stability for the data center.

Secondly, expanding existing transmission and distribution networks to supply gigawatt-scale power to data centers is a lengthy process. It involves navigating tough regulatory reviews and often facing local opposition. By comparison, deploying gas turbines is relatively straightforward.

According to U.S. tech outlet Wired, in Texas—a state where data centers are proliferating—officials are now authorizing the construction of 'standalone gas turbine power plants' dedicated solely to supplying power to data centers. In Tennessee’s Memphis, Elon Musk—founder of Tesla—is building the 'Colossus' AI data center, which features a separate power complex equipped with a staggering 69 large gas turbines. This mega-project was completed in just 122 days—a clear demonstration of the rush to expand capacity.

"A Choice Between AI and Electricity Bills Looms"

Storage facility of Henry Hub, the world's largest natural gas trading hub located in Louisiana, USA. CME Group website

Some have proposed using small modular reactors (SMRs) to supply electricity to data centers, but SMRs remain a nascent technology, with only a handful of countries having prototyped them. For now, gas turbines remain the dominant power source for data centers, and this trend is likely to continue at least until 2030.

If an AI-driven 'gas crisis' does materialize, there is a risk that the global shockwave will spread far beyond the United States. After all, the U.S. is both the world’s largest consumer and exporter of natural gas. Should Henry Hub’s natural gas price increase, LNG export prices would climb as well, pushing international fuel prices higher. Smith warns, "Ultimately, it’s the consumer who will suffer most. When electricity bills actually rise, we’ll be forced to choose whether to use that power for AI computing or to keep household electricity rates in check. Either way, it will be a terrible choice."

Energy supply turbulence driven by AI is rapidly emerging as a hot-button issue among American policymakers, as well. On July 27, the Trump administration publicly introduced the so-called 'Ratepayer Protection Pledge.' At its core, the pledge is aimed at preventing rising U.S. household electricity bills caused by AI data centers. Going forward, major tech companies building new hyperscale data centers will be required to invest in their own power plants and secure electricity independently.

This content was produced with the assistance of AI translation services.