Addressing the Surge in demand in Data Center Power fueled by Artificial Intelligence

I attended the AI + Energy Summit hosted by the Special Competitive Studies Project (SCSP) on September 26 in Washington DC along with senior policymakers, researchers, energy, artificial intelligence (AI) industry leaders, and other experts. The discussions centered around the vital role of near-term abundant and low-cost energy in maintaining U.S. competitive advantages in AI.

Artificial Intelligence is fueling a big boom in data centers and energy demand. In January, the International Energy Agency (IEA)?forecasted?that global data center electricity demand will more than double from 2022 to 2026, with AI playing a key role in that increase. Per Goldman Sachs Research estimates - at present, data centers worldwide consume 1-2% of overall power, but this percentage will likely rise to 3-4% by the end of the decade. Also, data center power demand will grow 160% by 2030.

Having extensive experience with data centers, each day as I drive past the massive data centers in Ashburn, Virginia (Widely known as the Data Center Capital of the World), I ponder about the impact of the growing demand for electricity in the U.S. that has not been seen in a generation. The MIT Technology Review (May 23, 2024) had a very catchy title in an article “AI is an energy hog ….” ?The big question is how do we prepare for this? What can we do?

It has been widely reported that it is ?extremely energy intensive to train AI models, which burn through power at a far higher rate than traditional data center activities. A ChatGPT query, for example, uses ten times more energy than a standard Google query, says David Porter, a vice president at the Electric Power Research Institute. Porter says that while 10-20% of data center energy in the U.S. is currently consumed by AI, that percentage will likely “increase significantly” going forward.?

Speaking at the conference - U.S. Secretary of Energy Jennifer Granholm highlighted the ambitious goal of a 100% clean energy grid by 2035 and the pivotal role that AI will play in our future! She assured the enthused audience that her agency is doing everything it can to prepare the country to meet?the rising demand?from energy-guzzling?AI data centers. Department of Energy (DOE) has been seeking to reassure utilities, Big Tech, and consumers that the agency can harness AI's unexpected and skyrocketing power needs. It?launched?the Frontiers in Artificial Intelligence for Science, Security and Technology (FASST) Initiative in July, positioning itself as a hub to research and find solutions to tackle emerging AI issues. Secretary Granholm spoke about the modeling done inside DOE and the need to double the size of U.S. electric grid, adding probably gigawatts of clean power on the grid to get to that goal and include the notion that there is growth in demand.

Former Google CEO Eric Schmidt lauded the huge improvements in healthcare, education, and science. He stated that energy constraints are a top concern for AI development in the U.S. and predicted a potential power shortage within four years. He stressed that there needs to be urgent collaboration with policymakers to prevent this bottleneck. He pointed out that the data centers need cost at base load power which is 3% of the electrical grid (in the U.S.) and there are forecasts that it will go to 7% or 8% in the next 5 to 10 years. China is preparing for that kind of growth.

Schmidt has stated in the past ?(in an Article: AI Revolution: Eric Schmidt’s Insights on the Future of Technology and Global Competition) that the race to develop these advanced AI systems is intensifying, and with that the gap between frontier models and others appears to be widening and this competition requires staggering amounts of capital and computing power, with estimates ranging from $10 billion to over $100 billion for cutting-edge systems.” Given the resource demands which are reshaping the competitive landscape, Schmidt suggested that only a handful of companies and countries may be able to compete at the highest levels.

The Panelists stressed that AI has the potential to revolutionize the energy industry, but it will require collaboration, investment, and innovative thinking to meet the challenges ahead. Rep. Don Beyer shared insights on the federal government's role, fusion energy advancements, and AI's potential in driving sustainable solutions and national security.

Robert Blue, President and CEO of Dominion Energy, shared insights on AI-driven demand, renewable energy targets, and regulatory challenges. The CEO pointed out that data centers represented?24 percent of Virginia Power's electricity sales for the years ended December 2023, up from 21 percent in 2022, and that Dominion is constructing a Coastal Virginia Offshore Wind commercial project, which when fully constructed in 2026, will deliver clean, renewable energy to the grid, thereby avoiding millions of tons of carbon dioxide emissions annually, and help Dominion’s goal of Net Zero carbon and methane emissions by 2050.

Senator Schatz emphasized the need to balance AI's growing energy demand with climate commitments, advocating for partnerships between tech companies and utilities to scale clean technologies.

Some key takeaways from the AI+ Energy Summit include the fact that the demand for AI infrastructure is unprecedent and it is time for sustainable energy solutions to support this growth. There needs to be close collaboration and partnerships between public and private sectors along with strategic investments. A reduction in AI power consumption presents a clear win-win opportunity for everyone!! It requires a multi stakeholder approach to balance AI’s energy use and emissions with its societal benefit.

On September 12, 2024, as part of the Biden-Harris Administration’s comprehensive strategy for responsible innovation, the White House had convened leaders from hyperscalers, artificial intelligence (AI) companies, datacenter operators, and utility companies to discuss steps to ensure the United States continues to lead the world in AI. Participants considered strategies to meet clean energy, permitting, and workforce requirements for developing large-scale AI datacenters and power infrastructure needed for advanced AI operations in the United States. To accelerate public-private collaboration in advancing U.S. leadership in AI, the Administration announced several new actions (more details on the White House website).

Thank you to The Special Competitive Studies Project (SCSP) for hosting a successful event and to all the amazing speakers!

Disclaimer: The opinions expressed in this article are the author’s own and has not been written in his capacity as a U.S. State Department Official.

?

?