Clean energy in the AI era becomes a hot cake: Google partners with Berkshire Hathaway's utility company for energy collaboration

According to the Zhitong Finance and Economics APP, Alphabet Inc's technology giant Google (GOOGL.US) has reached an important cooperation agreement with a utility company owned by the "Stock God" Buffett in Nevada, USA, and a startup focusing on geothermal energy, to use efficient and clean geothermal energy to power Google's large data centers. The thriving data centers, which currently focus on AI training/inference workloads, have significantly increased the electricity demand across the United States.

According to a statement released by Google, the energy startup Fervo Energy will develop a new geothermal power plant to provide up to 115 megawatts of power to NV Energy, a subsidiary of Warren Buffett's Berkshire Hathaway Energy company. NV Energy will then sell the electricity resources to Google, but this multi-party agreement still requires formal approval from national regulatory agencies.

Unlike solar and wind energy infrastructure, geothermal energy is not affected by weather and seasons, providing stable baseload power. Therefore, geothermal power plants can operate 24/7, mainly using water heated by underground rocks to generate electricity without generating a significant amount of carbon emissions.

Geothermal energy is the natural release of heat from the Earth's interior, a heat source that will last for billions of years throughout the Earth's lifecycle. From the Earth's perspective, geothermal energy is considered a renewable resource, and geothermal power generation processes emit almost no greenhouse gases such as carbon dioxide, making it a highly efficient clean energy source.

Geothermal energy originates from the Earth's high-temperature magma in the interior, transferred to the surface through crustal fractures or hot fluids. Fervo Energy, the energy startup company, uses horizontal drilling in geothermal reservoirs to increase electricity production.

Currently, large-scale data centers led by tech giants such as Google, Microsoft, and Amazon are consuming increasingly more electricity. This poses significant pressure on tech companies that have set goals to reduce carbon emissions causing global warming, prompting these companies to shift their focus towards clean and renewable resources for power generation. Statistics show that the electricity capacity Google has agreed to purchase from NV Energy and Fervo can power approximately 86,000 households.

In the AI era, clean energy supply becomes increasingly important

With ChatGPT sweeping the globe in 2023, the launch of the Sora WenSheng video large model in 2024, and NVIDIA's unparalleled performance in the AI field with "AI shovelers," it may indicate that human society will gradually enter the AI era starting in 2024. Data centers play a crucial role in the comprehensive rise of generative AI like ChatGPT. As the market demand for AI data centers accelerates, the already "power-hungry beasts" data centers are expected to see a surge in energy demand. The International Energy Agency predicts that by 2026, the total electricity consumption of global data centers will increase from approximately 460 terawatt-hours in 2022 to at least 1000 terawatt-hours. Goldman Sachs released a research report earlier this week, showing that by the end of 2030, AI will drive a staggering 160% increase in data center electricity demand. Goldman Sachs predicts that just to support AI data centers, U.S. electric utilities will need to invest approximately $50 billion in new generation capacity.

"On average, the electricity supply scale required for AI services such as ChatGPT queries and reasoning is almost 10 times that of normal Google searches. In this difference, how the U.S., Europe, and the entire world will consume electricity—and how much it will cost—is about to undergo a revolutionary change," Goldman Sachs stated.

"For years, data centers have had a very stable demand for electricity, even as their workloads increase. Now, with the slowing pace of improving power usage efficiency and the rise of the artificial intelligence revolution, we expect data center electricity demand to grow by 160% by 2030," added Goldman Sachs.

According to forecast data from the Boston Consulting Group, the share of data centers in electricity consumption in the U.S. alone is expected to double, from 126 terawatt-hours in 2022 to 390 terawatt-hours in 2030. Terawatt-hours are used to describe the largest power levels and are typically used for national energy statistics, planning, and evaluation of large energy projects. Large industrial facilities, such as super steel plants, may consume less than 10 terawatt-hours of electricity in a year.

Major data centers like Google and Microsoft have an incredibly strong demand for clean energy, mainly due to the global decarbonization trend, focusing on clean attributes such as wind power and geothermal energy as the most important sources of electricity generation, if not the only ones. The energy think tank Ember's latest statistics show that due to the continued rapid expansion of renewable energy sources such as wind and solar power compared to traditional fossil fuels like coal, the share of renewable energy in global electricity generation rose to a record 30% last year, and the think tank expects this trend to accelerate further this year