Meta data center in Arizona|Mesa Data Center|Facebook
Artificial intelligence is hungry for electricity, exponentially increasing data centers’ energy use and challenging power infrastructure in the country. And a conference last week addressed just that.
The CERAWeek by S&P Global is an annual energy meet that typically focuses on oil and gas deals and climate change goals. But this year, it focused on how AI is driving electricity demand.
What are they saying?
Amazon Web Service’s vice president of engineering pointed out that one new data center is added every three days globally, stressing the challenges the growing numbers of these huge buildings pose to power grids. The meeting also discussed how data centers are making it harder for companies to transition to cleaner energy sources.
Even Bill Gates said that the success of a data center is highly dependent on a reliable source of electricity, which tech giants like Microsoft, Meta and Amazon are constantly searching for as they need a cost-effective and continuous flow of electricity.
The International Energy Agency estimates that electricity consumption from data centers, AI, and cryptocurrency could double by 2026.
Utility companies, such as Dominion Energy and Southern Company, are revising their forecasts due to the expansion of these large facilities—expecting a 6,600 megawatts demand growth through the winter of 2030, 17 times greater than the previous forecast.
It is not good, considering the Biden administration aims to eliminate carbon emissions from the US electricity sector by 2035.
The surge in AI-driven power demand coincides with other factors stressing the grid, including increased electric power usage in manufacturing, transportation, heat, and other industries, says the Wall Street Journal.
Not just the US power grid
- Microsoft’s water usage in Des Moines prompted the city to pause future data center projects until the company reduces water consumption.
- Google data centers’ water usage in Oregon received public backlash for consuming millions of gallons a day.
Tech developers say the increase is due to the large amounts of water needed to cool the supercomputers that heat up when processing AI products like ChatGPT.
Overall, the convergence of AI’s electricity and water demand underscores the complexity of balancing energy needs with environmental sustainability goals.