Bill Gates Believes A.I. Can Solve Its Own Energy Problem

"Let's not go overboard," said Bill Gates of concerns regarding A.I.'s vast energy consumption.

Bill Gates is urging the world not to overreact when it comes to A.I.’s vast energy consumption. Despite concerns regarding the high electricity use of data centers required to power the new technology, the Microsoft (MSFT) co-founder claims this will be offset by innovations in A.I. regarding sustainable energy and carbon reduction.

“Let’s not go overboard,” said Gates while speaking yesterday (June 27) at the Breakthrough Energy Summit, as reported by the Financial Times. While Gates said data centers could globally increase electricity usage by 2 to 6 percent, the billionaire believes tech solutions will act as a countervailing force. “The question is, will A.I. accelerate a more than 6 percent reduction? And the answer is: certainly.”

Amid a race for A.I. dominance in the world of Big Tech, companies like Microsoft, Meta (META), Google (GOOGL), Amazon (AMZN) and OpenAI are funneling billions into the development of data centers across the U.S. to train and deploy A.I. models. But with this rapid growth comes soaring energy use. A query on OpenAI’s ChatGPT, for example, takes up nearly 10 times as much electricity to process as a search on Google, according to a recent report from Goldman Sachs (GS). And while data centers currently consume 1 percent to 2 percent of global power, this will likely rise to 3 percent to 4 percent by 2030 with carbon dioxide emissions from data centers possibly doubling as of 2030, the report predicts.

While it is difficult to accurately predict the energy consumption of “these data centers in 10, 20 years from now,” the demand for A.I. “is going to grow very fast,” Fengqi You, an energy engineering professor at Cornell University, told Observer, noting that the current capabilities of U.S. infrastructure are an especially pertinent issue. The development of data centers across the country have already put a strain on regional power grids.

Gates is a prominent believer in technology’s abilities to solve climate issues. He’s the founder of Breakthrough Energy Ventures, a climate-focused investment firm that has committed more than $3.5 billion to tech startups cutting greenhouse gas emissions, and has given more than $1 billion to the nuclear energy startup Terrapower.

How close is A.I.’s green electricity potential?

According to Gates, major tech companies are willing to pay a “green premium” to develop data centers with new sources of power. Amazon, for example, has claimed it will run its data centers entirely on green energy by 2025, while Google and Microsoft say they can accomplish this by 2030.

“They’re all paying attention to it, and they want to go down that path,” Baratunde Cola, the founder of carbon nanotude company Carbice Corporation, told Observer of Big Tech’s goal towards cleaner energy. But what’s more unclear is “how fast we can go there,” he added.

It will take time to untap the amount of green electricity needed for such a transition, according to Gates, who noted that global goals to reach net zero emissions by 2050 may be unachievable. “Another 10 or 15 years might be more realistic,” he said.

Gates isn’t the only prominent tech leader betting on A.I.’s green possibilities. In April, Amazon’s Jeff Bezos invested $100 million via his environmentally-focused Bezos Earth Fund into an initiative encouraging proposals to use A.I. to combat climate change. And Geoffrey Hinton, an A.I. academic who previously worked for Google, earlier this month joined the advisory board of CuspAI, a startup aiming to create A.I.-generated materials used in carbon capture and storage to reduce greenhouse gas emissions.