Executives from OpenAI, Anthropic, Google and Microsoft attended a meeting at the White House on Sept. 12 to discuss building robust energy infrastructure for artificial intelligence and high-performance computing needs.
Representatives for the companies — including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei — argued that America’s aging energy infrastructure could not adequately service the massive energy needs of the budding AI sector.
Investment banking firm Goldman Sachs estimates AI will have seen a “160% increase in data center power demand” by 2030. Spokespeople for OpenAI asserted that building AI-grade energy infrastructure will create an influx of jobs inside the United States.
Data center power demand 2022-2030. Source: Goldman Sachs
OpenAI personnel claimed that building this energy infrastructure will also contribute to the geographic decentralization of AI and ensure that the US remains on the cutting edge of innovation.
Executives at the meeting also explained that failing to bolster the current energy infrastructure and falling behind in the AI race poses a national security risk to the US.
Related: California lawmakers push through controversial AI bill opposed by OpenAI
Meeting massive energy demand from AI and high-performance computing
Phil Harvey, CEO of Sabre56 — a data center consulting company — recently told Cointelegraph that AI data centers cost between $3 million and $5 million per megawatt to operate. This assessment was corroborated by a report from the International Energy Agency (IEA), which found that a typical ChatGPT search consumes 10 times the amount of energy required for a standard Google search.
Workload demand for data centers 2015-2023. Source: Goldman Sachs
Indiana lawmakers recognize this pressing need and have assured corporate giants like Google, Meta and Microsoft that the state can deliver massive quantities of outage-free energy to power data center operations.
Indiana Secretary of Commerce David Rosenberg stressed that the state had “ample” amounts of energy and water, which are needed to cool down data centers — to meet industrial demand.
The state’s rich natural resources have spurred a $14-billion investment into data center facilities by Big Tech companies seeking a home for robust AI and high-performance computing operations.
Magazine: AI may already use more power than Bitcoin — and it threatens Bitcoin mining