AI energy needs could be a bottleneck for the US, Eric Schmidt says

The post AI energy needs could be a bottleneck for the US, Eric Schmidt says appeared first on Android Headlines.

Apr 15, 2025 - 15:44

Former Google CEO Eric Schmidt is becoming an increasingly prominent figure in the artificial intelligence segment. The tech expert has been very vocal with his views on AI development. Occasionally, his statements have sparked controversy due to their somewhat pessimistic tone. Now, Schmidt is warning the US government about energy needs as a potential bottleneck to dominating the AI segment.

Eric Schmidt appeared before the US House Committee on Energy and Commerce to speak about the future of AI. He focused specifically on one area: energy generation. Schmidt believes that the US will soon reach a point where energy generation capabilities will limit the development of artificial intelligence. This could give China an advantage in the future. He also highlighted that current US regulations prevent the rapid set of new, efficient approaches to generating energy to meet AI needs.

AI energy needs could exceed the US’s power generation capacity, former Google CEO warns

“We need energy in all forms, renewable, non-renewable, whatever. It needs to be there, and it needs to be there quickly,” the former Google CEO told the lawmakers. “People are planning 10 gigawatt data centers now,” he stated. “An average nuclear power plant in the United States is one gigawatt,” he added later. Overall, Schmidt questioned the current capacity of the US’s electrical infrastructure to keep up with the speed of AI development.

Schmidt’s words point to an impending AI bottleneck in the US. To avoid this, he believes the country must invest in new approaches to energy generation. It also needs to rethink how the electrical grid is managed and the overall use of available power.

Regarding China, Schmidt warned of the risk of it achieving “superintelligence” before the US. The first country to achieve this will likely be the immediate winner in the AI race. “If [China] becomes superintelligible first, it changes the dynamics of power globally, in ways that we have no way of understanding or predicting,” he said.

SMRs are “the right answer,” but regulations prevent them from being set up quickly

Schmidt didn’t limit himself to telling lawmakers to work on alternative energy production. The tech expert even offered his perspective on the ideal alternative. He said that small modular reactors (SMRs) are “the right answer.” However, he expressed frustration with the regulations affecting the set of these types of plants. “One of my personal frustrations is the regulatory structure around nuclear and SMRs… How many SMRs are in use in America today? Zero,” he said. He also noted that a “fast approval time is considered to be 12 years. That defies logic.”

AI energy needs are expected to increase notably in the coming years

By 2030, the International Energy Agency (IEA) projects that data centers will double their energy consumption. In particular, AI data centers will quadruple their energy needs. AI’s power requirements could soon surpass those of industries like steel and cement manufacturing, The Guardian reports.

US representatives project that, by 2028, AI energy consumption will account for 12% of total US consumption. For reference, the figure was 4.4% in 2023. These projections suggest that Schmidt’s recent warnings are not far-fetched.

The post AI energy needs could be a bottleneck for the US, Eric Schmidt says appeared first on Android Headlines.