When it comes to U.S. electricity demand, chatbots matter more than cars

Artificial intelligence is set to become a game changer for the electric power industry.

Whether or not you enjoy sitting in front of a computer and conversing with a human-seeming artificial intelligence (AI) platform (i.e., a chatbot), the AI boom will, if nothing else, affect your energy bill. AI is set to become a game changer for the electric power industry – especially, though not solely, in the United States. Principally as a result of AI, U.S. electricity demand is about to start posting meaningful growth for the first time in two decades. Enabling the power grid to manage this growth in demand, while simultaneously shutting down aging coal-fired power plants, will require an all- of-the-above strategy. At Raymond James, analysis of this trend has involved a collaborative, cross-industry effort between the technology and energy research teams.

Demand for electricity set to skyrocket

It may come as a surprise to some of our readers that U.S. electricity demand has been flattish over the past quarter-century – even as population and GDP have continued to grow. The reason is energy efficiency: everything from lightbulbs to air conditioners is more efficient than the older equipment being replaced. U.S. electricity demand in 2023 was up only 10% from 2000, equating to average growth of only 0.4% per year. AI is about to change that in a big way. We forecast that US electricity demand will grow at an average of 2.8% per year through 2030, with AI-related demand comprising two-thirds of incremental demand. The other one-third comprises electric vehicles and, well, everything else! Data centers supporting the AI boom are extremely energy-intensive, even more so than data centers that have been around since the early days of the Internet. Simply put, whenever someone engages with a chatbot – and this is happening countless times every day – a hefty amount of electricity is used. Furthermore, data centers provide a textbook example of mission-critical electricity users: they cannot afford to lose power even for a minute. Data centers run by the large third- party providers and hyperscale cloud companies typically have sufficient on-site generators and temporary battery backup to remain operational during grid outages.

So, is this good news or bad news for the electric power industry?

The short answer is: good. Utility companies obviously want to sell more electricity to their customers, and as mentioned earlier, they have had very little organic growth (on the whole) for a long time. That said, utilities also face the risk of managing a power grid that is unprepared for the increasing demand. In the worst-case scenario, it could lead to systematic load shedding- deliberate power outages affecting a large proportion of the population- as a way of reducing the stress on the grid. We are not the only ones who see the risks. In its latest Report Card for America’s Infrastructure, the American Society of Civil Engineers gave electric power infrastructure a lackluster C–grade and forecasted an investment shortfall of $200 billion by 2029.

Power grids in depth

An important but not always understood feature of the U.S. electric grid is the extent of its geographic bottlenecks. The two major grids – Eastern and Western Interconnections – have minimal connectivity, for reasons that go back to the first half of the 20th century. As it stands, there is very little transfer capacity between them. Data center operators in the east need to rely upon physical electricity supply that is also in the east. Likewise, coal plants being shut down in the Rocky Mountain region will be mostly replaced by the wind and solar projects located in the west. Texas, the ERCOT grid, is largely isolated from both the Eastern and Western Interconnections. This presented a major problem during the snowstorm in early 2021, when, despite sky-high power prices, Texas was unable to source urgently needed electricity from its neighbors. In the long run, as part of infrastructure upgrading/ modernization across the board, what needs to happen is for connectivity between the regional grids to be greatly bolstered.

A marathon, not a sprint

What we have discussed in this article encompasses a multi-decade story. In the electric power industry, nothing changes on the spur of the moment – in other words, do not expect any immediate transformation. Utility management teams and policymakers alike are aware of what needs to be done, but complex infrastructure projects are always prone to delays. The capital to make these investments is available, and so is the technology, though labor can be a constraint. This is a proverbial marathon rather than a sprint, which means that a wide range of companies stand to benefit from these opportunities for decades to come.

Click here to read the entire article.

Material created by Raymond James for use by its financial advisors. All expressions of opinion reflect the judgment of the authors and are subject to change. Past performance may not be indicative of future results. There is no assurance any of the trends mentioned will continue or forecasts will occur. The performance mentioned does not include fees and charges which would reduce an investor’s return. Dividends are not guaranteed and will fluctuate. Investing involves risk including the possible loss of capital. Asset allocation and diversification do not guarantee a profit nor protect against loss. The companies engaged in business related to a specific sector are subject to fierce competition and their products and services may be subject to rapid obsolescence. There are additional risks associated with investing in an individual sector, including limited diversification.