TL;DR
Get backup power and energy gear delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Joseph Santamaria, Amazon Web Services’ general manager for energy and utilities, argues that AI tools can help utilities speed up grid interconnection studies and address project backlogs. He cites examples from Duke Energy and Pacific Gas & Electric, but the scale of any systemwide time savings and the role AI can play in clearing queues remain uncertain.
Joseph Santamaria, Amazon Web Services’ general manager for energy and utilities, says utilities can use AI-powered grid tools to speed up studies that determine whether proposed power projects can connect safely to the U.S. electricity network. His argument comes as projects face years-long waits for review, though the examples he cites do not establish how quickly AI could reduce backlogs across the industry.
In a commentary published by Utility Dive, Santamaria said more than 2,000 gigawatts of potential power projects in the United States are waiting to be reviewed. Those reviews, known as interconnection studies, assess how a proposed generator or other power resource would interact with existing lines, substations and generators. Utilities and grid operators use the results to determine whether a connection is safe and what upgrades may be needed.
Santamaria said an interconnection study can take months, while projects may wait years to reach the front of the review line. He put the average wait at about two years in the early 2000s and said it has grown to five years and counting amid rising demand. The commentary did not specify the dataset or methodology behind those wait-time figures.
According to Santamaria, AI tools can run thousands of grid simulations across different scenarios, rather than having engineers test a smaller set of cases sequentially. He said that could allow some studies to be completed in hours instead of months. Under the approach he described, software generates analyses and recommendations, while human engineers make final decisions.
Faster Reviews Could Ease Grid Queues
Long waits for interconnection studies can delay power projects even when developers are ready to build, leaving planned generation unable to connect until technical reviews and any required upgrades are addressed. If AI tools help utilities complete those reviews more quickly, they could make it easier to assess proposed resources as electricity demand rises from electrification, industrial activity and data centers.
The potential effects extend beyond project developers. Santamaria argues that faster planning can help utilities meet demand while maintaining reliability and affordability. But AI analysis cannot by itself build transmission lines, approve projects or settle cost disputes. Its value depends on utilities adopting the tools, having usable grid data and retaining engineers and regulators to review the results.
The issue also has a customer-cost dimension. New infrastructure may be needed to serve large power users, and critics have questioned whether those costs could be passed to existing customers. Santamaria says outcomes depend on how projects are structured and who pays for upgrades. The claims about customer savings and affordability should be read as his argument, not a guarantee that every new project will lower bills.
Utility Examples Cited by AWS
Santamaria pointed to Duke Energy, which he described as the country’s largest utility, as an example of AI-assisted grid planning. He said the company reduced study time while doubling the number of projects evaluated in a year. The commentary did not give the precise period, project count or size of the time reduction.
He also cited Pacific Gas & Electric, which brought data from more than 80 sources across 20 systems into a planning tool, according to the commentary. Santamaria said the program helped prioritize more than 17,000 grid upgrades and cut costs by 20% within a $100 million pilot. He also attributed millions of customer outage minutes eliminated to the program; the source did not provide a timeframe or further measurement details for that figure.
Regulatory interest is another part of the backdrop. Santamaria said the Federal Energy Regulatory Commission brought utilities, technology companies, government agencies and other participants together this summer to discuss AI and grid planning. That discussion indicates attention to the issue, but the source does not say FERC endorsed particular tools or adopted a new requirement.
“These systems run thousands of simulations to generate recommendations while human engineers make the final decision.”
— Joseph Santamaria, describing AI-assisted grid studies
Evidence and Rollout Remain Open
The commentary does not provide independent evaluations of the Duke Energy and PG&E examples, detailed baselines for the reported improvements, or evidence that the same results can be achieved across utilities. It also does not identify which AI systems were used, how their performance was measured, or whether study times fell consistently across different types of projects.
It remains unclear how much of the reported queue could be cleared by faster studies rather than by other constraints, such as limited engineering staff, required grid upgrades, permitting or changes to project proposals. The source gives a national total for potential projects awaiting review but does not break it down by project type, location or stage in the process.
Customer costs are also unresolved on a project-by-project basis. Santamaria cites an Amazon agreement with Northern Indiana Public Service Co. expected to save existing customers about $1 billion over 15 years, with Amazon covering costs associated with new data centers. That is a projected result for one arrangement, not proof that other large-load agreements will have the same terms or outcomes.
Utilities and Regulators Weigh Adoption
The immediate next step is whether utilities expand AI-assisted planning beyond pilots and use it in routine interconnection reviews. The commentary does not announce a new AWS product, utility deployment or regulatory decision, and it gives no timeline for wider adoption.
Further evidence would help establish whether these tools reduce total queue times, rather than only speeding up individual calculations. That includes independently reported results, clear comparisons with prior review timelines and information about how utilities validate recommendations and account for required grid upgrades. FERC’s discussion this summer may form part of ongoing industry attention, but no specific follow-up action is detailed in the source.
For customers and project developers, the key developments to watch are changes in study times, the number of projects utilities can evaluate, infrastructure spending and the terms that assign costs to new power users. Until those outcomes are reported across more deployments, AI’s potential to relieve grid backlogs remains a claim supported by selected examples rather than a proven systemwide solution.
Key Questions
What are interconnection studies?
Interconnection studies assess how a proposed power project would affect the electricity grid and what changes may be needed to connect it safely. The reviews examine interactions with equipment such as power lines, substations and generators.
How does Santamaria say AI could help?
Santamaria says AI tools can run thousands of simulations across different grid scenarios and produce recommendations for engineers to review. He argues this could shorten some study work from months to hours, but the source does not establish that result for the industry as a whole.
Have utilities already reported results?
Santamaria cites reported work by Duke Energy and Pacific Gas & Electric, including faster study work at Duke and a PG&E planning program that prioritized grid upgrades. The commentary provides limited detail on methods and baselines, and does not include independent verification of the figures.
Will AI clear the project backlog by itself?
No such outcome is established in the source. Faster studies may address one part of the process, while grid construction, permitting, staffing and project changes can also affect how quickly power projects connect.
Could new data centers increase existing customers’ bills?
The source says the effect depends on project arrangements and who pays for new infrastructure. Santamaria cites one Amazon–Northern Indiana Public Service Co. agreement with projected customer savings, but that example does not establish how other projects will affect bills.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
