Analyzing the latest data on the U.S. energy sector, I am recording an unprecedented shift: the announced capacity of gas-fired power plants aimed at directly powering data centers has surged from 97 GW at the end of 2025 to more than 189 GW by mid-2026. This jump is a direct consequence of the artificial intelligence boom, which is redrawing the country's energy landscape.
For context: at the beginning of 2024, this figure barely exceeded 4 GW. Thus, over two and a half years, the volume of targeted projects has increased more than 40-fold, and since December alone—nearly doubled. My sample, based on global monitoring data, includes both announced initiatives and facilities at the pre-construction stage, as well as plants already under construction. It is important to emphasize: a significant portion of these megawatts may never materialize beyond paper.
The Era of Autonomous Energy Centers
The key driver is the spread of the behind-the-meter model. This involves building power plants directly at the data center site, allowing operators to bypass congested and slow power grids, where connection wait times stretch for years. This is no longer a trend but a new reality: the expansion of gas-fired generation in the U.S. is now inextricably linked to the appetites of the AI industry.
As early as January, I estimated the total portfolio of U.S. gas projects at nearly 252 GW, of which more than a third (about 97 GW) were intended specifically for data centers. If this entire volume were realized, installed capacity would grow by nearly 50%, with capital expenditures exceeding $416 billion. The new figures show that it is precisely the "data center" segment of the market that continues to expand exponentially.
Political Momentum and the Bet on Gas
This shift toward autonomy coincided with a change in the White House's course. In March, the largest players—Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI—committed to self-financing new energy capacity and grid infrastructure. The goal is to avoid shifting the burden of growing consumption onto ordinary consumers. By July, more than 200 additional organizations, including utility companies and state authorities, had joined this initiative.
However, the bet is placed primarily on natural gas. This allows for faster commissioning of capacity but creates long-term risks. I would advise against treating 189 GW as a guaranteed forecast. Realization depends on financing, permits, public opposition, and, critically, on a shortage of gas turbines—manufacturers' order books are already booked through 2030, and for two-thirds of the projects, equipment has not even been specified.
Energy Geopolitics: The U.S. vs. China
This gas boom has allowed the U.S. to overtake China in the total capacity of gas projects under development (252 GW versus 153 GW). But the approaches differ radically. China is betting on locating data centers in regions with surplus solar and hydroelectric power, following a state policy of energy independence. The U.S., by contrast, chooses speed, gaining a stable source of capacity at the cost of additional emissions and the risk of creating excess infrastructure if AI-driven demand fails to meet expectations.
The epicenter of the boom is Texas. In January, I recorded 80.6 GW of gas capacity under development there, of which about 40 GW is directly linked to data centers. The state, which already hosts around 300 data centers with another 200 in development, actively leverages streamlined approval procedures. A telling example is Amazon, which purchased a site in Pecos County for a campus with its own gas-fired plant, GW Ranch, with a capacity of 7.65 GW and 35 turbines—operating before connection to the main grid.
My conclusion: we are witnessing the formation of a new energy paradigm where AI becomes the primary driver of demand. However, betting on fossil fuels in the long term is a double-edged sword: rapid growth now could turn into serious environmental and economic costs if the AI investment bubble begins to deflate. The market is clearly underestimating these risks, focusing on short-term gains.