The size and energy demand of data centers in the United States is growing rapidly, now comparable to the electricity consumption of entire states. One example is Meta’s Hyperion artificial intelligence data center. When completed, the center’s electricity demand is expected to be the same as the state of South Dakota.
The company announced last week that it would finance the construction of seven additional natural gas-fired power plants to power the $27 billion project. Together with the three previously planned plants, the 10 plants to be built in Louisiana will have a total capacity of about 7.5 gigawatts. That’s more than the total power generation of some U.S. states.
Meta has for years presented itself as an environmentally responsible company, publishing sustainability reports and emphasizing its investments in renewable energy projects. The company even has long-term nuclear power contracts. But the Hyperion project raises questions about the real nature of these commitments.
Natural gas has long been touted as a “transitional fuel” — a temporary solution until renewable energy technologies are developed. Meta is also believed to have justified its decision internally. However, experts say that this argument is already outdated: while the cost of solar energy and batteries has fallen sharply, the cost of gas turbines has increased.
According to calculations, these power plants in Louisiana will emit about 12.4 million tons of CO₂ annually. This is about 50% more than Meta’s total carbon footprint recorded in 2024. Moreover, this figure does not include methane leaks from the extraction and transportation of natural gas.
Methane gas is particularly dangerous for the climate — it has a warming effect 84 times stronger than carbon dioxide. The level of leakage along the gas infrastructure in the United States is reported to be about 3%, which calls into question the presentation of natural gas as a “clean energy”.
Meta has not yet made an official statement on the issue. The company may rely on carbon credits to offset its carbon emissions in the future. But experts say more transparent and realistic reporting is needed to offset such large emissions.
So the race between tech giants for artificial intelligence and data infrastructure raises new questions about energy policy and climate commitments.
