OpenAI and Nvidia executives discuss AI infrastructure needs with Biden officials

OpenAI Chief Executive Sam Altman and Nvidia Corp. CEO Jensen Huang met with senior Biden administration officials and other industry leaders at the White House on Thursday to discuss how to meet the massive infrastructure needs for the projects of artificial intelligence.

On the tech side, attendees included Anthropic CEO Dario Amodei, Google President Ruth Porat, and Microsoft Corp. President. Brad Smith, according to people familiar with the meeting, which also had representatives from the energy sector. Government officials included Commerce Secretary Gina Raimondo, National Security Adviser Jake Sullivan and Energy Secretary Jennifer Granholm, according to the people.

The goal, according to a White House official, was to spur public-private partnerships around the development of AI data centers in the US. Topics included permitting, manpower, energy requirements and the economic impact of the facilities, people familiar with the meeting said.

OpenAI, for example, plans to spend tens of billions of dollars on an in-house AI infrastructure spanning data centers, power and transmission capacity, and semiconductor manufacturing — with investments from around the world. Company executives have been meeting with government officials for months about a range of issues related to the initiative, including national security concerns that could be associated with foreign capital.

“OpenAI believes that infrastructure is destiny and that building additional infrastructure in the US is critical to the country’s industrial policy and economic future,” OpenAI said in a statement Thursday. The company highlighted the economic benefits of investing in US data center projects, including a possible 40,000 jobs in a number of US states. OpenAI pointed to similar investments by China, which aims to become a global leader in artificial intelligence by the end of the decade.

Porat saw robust US energy infrastructure as crucial to ensuring US leadership in the emerging field of AI. “Today’s meeting at the White House was an important opportunity to advance the work needed to modernize and expand the capacity of America’s energy grid,” she said in a statement.

Anthropic and Microsoft declined to comment.

The AI-fueled surge in U.S. data center construction coincides with broader manufacturing growth spurred by the Chips and Science Act and the Inflation Relief Act — the signature semiconductor and clean energy subsidy programs passed in 2022 under President Joe Biden.

Those investments, along with data center expansion and other factors, are expected to increase electricity demand by 15 percent to 20 percent over the next decade, according to the Energy Department. Data centers could consume as much as 9 percent of U.S. electricity generation annually by 2030, up from 4 percent of total load in 2023, according to a May report by the nonprofit Electric Power Research Institute.

The Biden administration has said that renewable sources such as wind and solar, as well as battery storage and energy efficiency gains, are some of the best ways to meet the growing energy demand of data centers because they are rapidly scalable and are cost competitive.

“The near-term increase in data center-driven electricity demand is an opportunity to accelerate the development of clean energy solutions, improve demand flexibility and modernize the grid while maintaining affordability,” the Department of Energy said in a blog post last month.

However, the agency, which is due to publish an assessment of data center energy consumption by the end of the year, cautioned that forecasts of electricity demand growth “continue to evolve due to evolving use cases” and other factors.

Photo: Photographer: Graeme Sloan/Bloomberg

Copyright 2024 Bloomberg.

TOPICS
InsurTech Data-driven artificial intelligence