This audio is auto-generated. Please let us know if you have feedback.
Joseph Santamaria is general manager of energy and utilities at Amazon Web Services.
The U.S. economy is facing a looming shortage of electricity on the grid. With energy demand projected to grow more than 50% by 2050, utilities are under growing pressure to keep up.
The problem isn’t a lack of new energy projects. In the United States, more than 2,000 GW of potential power projects are waiting to be reviewed — enough to power every home in the country more than twice over. The challenge is bringing those projects from planning to execution.
Joseph Santamaria
Before utilities can bring new sources online, they must complete interconnection studies. These technical reviews determine how and where new power sources can safely plug into the existing grid. Each study takes months to complete, and projects can wait years just to reach the front of the line. In the early 2000s, the average project waited about two years. Today, surging power demand has more than doubled the wait to five years and counting.
New AI-powered grid tools can help shorten the time to approval by completing these studies in hours instead of months. With power demand growing rapidly, using new technology to speed up the approval process is critical to fueling economic growth, meeting demand from essential industries, and keeping electricity affordable and reliable for everyone.
Where a traditional study might test a handful of what-if scenarios one at a time, AI tools can model thousands of them at once by testing how new projects interact with the existing grid, including its power lines, substations and generators, across multiple scenarios. These systems run thousands of simulations to generate recommendations while human engineers make the final decision. With machines performing the legwork, operators can evaluate new projects much more quickly.
Utilities that have invested in AI grid tools are already seeing results.
Duke Energy, the largest utility in the country, has used AI-powered grid planning to reduce study time while doubling the number of projects evaluated in a year.
On the West Coast, Pacific Gas & Electric pulled together data from more than 80 sources across 20 different systems into a single planning tool, helping prioritize more than 17,000 upgrades across the grid. The program cut costs by 20% within a $100 million pilot program and eliminated millions of customer outage minutes.
Regulators are paying attention, too. This summer, the Federal Energy Regulatory Commission brought together utilities, tech companies, government agencies and others to discuss how AI can speed up grid planning. It's a clear sign that regulators see AI as a potential solution to grid backlog.
The urgency is clear. Increased demand from across the electric grid, from electrification to renewable energy projects to AI data centers, is adding pressure to an already backlogged system. Critics have raised questions about whether new facilities will drive up costs for existing customers or worsen the approval bottleneck.
The answer depends on how these projects are structured. When large energy users cover the cost of new infrastructure and utilities use modern planning tools, everyone benefits. In Indiana, an agreement between Amazon and Northern Indiana Public Service Co. is expected to deliver approximately $1 billion in savings to existing customers over 15 years, with Amazon covering all costs associated with building new data centers. Research from the Electric Power Research Institute confirms that with the right planning and pricing rules, new demand can help keep electricity prices in check.
These deals work best when utilities can evaluate and approve projects quickly enough to meet demand. That's where AI becomes a key player not just as a driver of new demand, but as the tool that makes it possible to bring new power demand onto the grid while shortening the time that everyone waits in line.
The utilities moving first on AI-powered grid planning are delivering real results — faster approvals, fewer outages and lower costs. For the communities where this new energy infrastructure is being built, that means reliable power and stable electric bills. The longer utilities stick with outdated planning processes, the longer families and businesses wait for the energy they need.