Big Tech has been one of the biggest and most innovative promoters of clean energy for decades. Silicon Valley’s biggest companies have poured significant amounts of money and effort into making themselves greener and supporting the growth of renewable energy. But in recent years, the conversation about the tech industry’s energy use has shifted from one of pride and optimism to a far more anxious conversation colored by an uncertain future. problem? artificial intelligence.
Renewable energy production is on the rise, and the United States has more incentive than ever to expand capacity, but that additional energy is dwarfed by the rapid increase in energy consumption driven by artificial intelligence. The training process for ChatGPT’s predecessor, GPT-3, is estimated to have required approximately 1,287 megawatt-hours of power and 10,000 computer chips. This is enough energy to power his 121 homes in the United States for one year and generate 550 tons of carbon dioxide.
As AI becomes more pervasive and the way the technology is designed and operated changes, the energy and carbon footprint of AI and associated data centers has grown exponentially, and this trend will only continue to grow. Currently, AI alone consumes as much energy in a year as the entirety of some developed countries, and rivals the energy footprint of Bitcoin. Consulting firm Gartner predicts that if business continues as usual, he will be solely responsible for a staggering 3.5% of the world’s electricity consumption by 2030.
Furthermore, all of that energy consumption is correlated with significant amounts of greenhouse gas emissions. Renewable energy accounts for a portion of Big Tech’s energy consumption, but this sector is unlikely to supply all of its energy needs. According to statistics from the International Energy Agency, data centers and power grids emit as much carbon dioxide each year as all of Brazil.
Indeed, the technology sector’s energy consumption, predicted largely due to breakneck growth in AI, is so large that Big Tech leaders are wondering whether they can secure enough clean energy to curb carbon emissions. Don’t just worry about having enough energy, period. “New data centers can be built faster than new power generation, but there is already a shortage,” the Wall Street Journal recently reported.
According to Bill Bass, vice president of engineering at Amazon Web Services, a new data center comes online somewhere in the world every three days. But without energy to power the new center, that schedule is not sustainable. Already, power supply issues have pushed many data center construction schedules from two years to six years.
This puts the technology industry in a huge predicament. They want to advance their ventures in clean energy, but they don’t have to wait. Toby Rice, CEO of giant US natural gas producer EQT, told the Wall Street Journal that “the technology industry is not going to wait seven to 10 years to build this infrastructure.” Ta. “That leaves us with natural gas.”
But other industry players are betting big on nuclear energy to meet the tech industry’s huge energy needs without hindering decarbonization goals. Others are looking to innovation and technological advances to save the sector from itself. “There’s no way we’re going to get there without a breakthrough,” OpenAI CEO Sam Altman said at a Bloomberg event on the sidelines of the World Economic Forum’s annual meeting earlier this year. . Altman believes this breakthrough will come in the form of nuclear fusion.
In 2021, Altman personally provided $375 million in funding to Helion Energy, a private U.S. fusion company that later signed a contract to supply energy to Microsoft. Mr. Altman is also chairman of the board of Oklo, a micro-reactor company that could go public this year. “Future AI systems will require enormous amounts of energy, and this splitting and merging will help make that happen,” Altman said.
Written by Haley Zaremba, Oilprice.com
Other top articles on Oilprice.com:
