Work continues on the new QTS data center in Eagle Mountain, Utah, on April 17, 2026. Today’s data center buildout has become the largest industrial boom since the railroad era of the 19th century, with $10 trillion in investment and almost twice the economic impact. (Photo by Spenser Heaps for Utah News Dispatch)
The sheer size of investment in data centers is drawing concern as it surpasses all other historical booms and crowds out other more job-creating ventures.
Until now, no industry had driven the U.S. economy with such force since the railroad expansion of the late 1800s, when titans battled one another to lay the transcontinental tracks that would connect America from coast to coast.
Railroads in the 19th century consumed about $173 billion in today’s dollars and created more than 2% of the nation’s annual economic output (GDP) over 20 years.
Today’s data center buildout will dwarf that amount, with $10 trillion in investment, and now have almost twice the economic impact — nearly 4% of GDP. That’s according to a report created by a Columbia Business School professor, Stijn Van Nieuwerburgh.
States are starting to feel left out of the windfall, many having granted tax exemptions before the boom.
Years ago, before artificial intelligence prompted a tsunami of building, states typically welcomed data centers supporting internet traffic, offering tax abatements and other perks. Many officials now are rethinking that approach as tax revenues take a bigger hit and high energy demands take a toll on affordability.
Nevada, for instance, trimmed tax abatements and will ask data centers to agree to foot the bill for energy and water demands, in a September executive order from Republican Gov. Joe Lombardo.
States where data center costs will be concentrated over the three years are Texas ($638 billion), Virginia ($249 billion), Ohio ($169 billion) and Georgia ($160 billion), according to projections from Industrial Info Resources, a market intelligence firm.
Southern states dominate the landscape for the number of data centers planned, with large numbers also proposed in Arizona, California, Florida, Indiana, Iowa, Indiana, North Carolina, Oregon, Pennsylvania and Washington state, according to a Pew Research Center report in April.
The profitability of AI has yet to catch up with the expense of the buildout.
“It is not a pretty picture right now,” said IIR’s senior vice president of research, David Pickering. “I think it will become profitable once things settle down a bit and more legislation is put in place to regulate how we handle AI.
“The big players will no doubt find a way to make it profitable,” he said.
There are signs the staggering cost of data centers is crowding out other investments such as housing.
Over the past two years new construction value for data centers has more than doubled to $85 billion annually, while all other construction value has dropped 8%, according to a Stateline analysis of U.S. Census Bureau data as of August.
As new data center value grew by $48 billion since 2024, new houses and apartments dropped by $29 billion, according to the data.
Adam Shapiro, a vice president at the Federal Reserve Bank of San Francisco, noted the change in a September social media post: “We’re seeing a pivotal shift in the US economy: investment is shifting away from residential investment and towards computers.”
Some, like the economist Paul Krugman, see the shift as hurting employment in the long run, despite more current hiring for construction workers.
“Because data centers employ so few workers, the vast spending on AI currently taking place will also significantly reduce the demand for labor,” Krugman wrote in an Oct. 11 analysis.
Josh Bivens, chief economist for the left-leaning Economic Policy Institute, agreed that the economy’s recent dependence on data centers and other AI spending is unhealthy, calling it “a narrow base of growth, basically one sector, and a pretty fragile one.”
“Anything that upsets expectations about the future potential profitability of AI investments will likely lead to pretty sharp cutbacks,” Bivens said. “I’m uneasy about how durable this strength might turn out to be.”
As for the historic nature of the buildout, Mark Higgins, author of a book on U.S. financial history, said there are parallels to the age of railroad building in the late 19th century.
“Railroads were a huge investment,” Higgins said. “In the 1800’s most of the capital was foreign, and it would attract all this money from Europe, then everything would crash — there were several panics —and Americans would buy the distressed railroads and profit, because they knew which were the good ones.”
Railroad construction, which did not include trains themselves, cost $4.3 billion between 1870 and 1890, about $173 billion in today’s dollars, according to historical figures provided to Stateline by Van Nieuwerburgh and adjusted for inflation.
Other large-scale construction booms consumed only about 1% or less of annual GDP: Canals in the early 19th century, electrification in the early 20th century, highways in the mid- to late-20th century, and telecoms and fiber between 1996 and 2003.
In a report this month the International Monetary Fund warned of the risk of lower stock prices if high expectations for AI fail to materialize.
“Although core firms in the AI ecosystem have been profitable so far … weaker or less durable investment and profitability in AI could trigger a correction,” the report warned.
The buildout will also double electricity consumption, Van Nieuwerburgh estimated, a continual thorn for data center projects that need more power to process vast data troves to power AI applications.
An Oregon report last month, which led to a continued freeze on data center approvals in the state, said the single most frequent concern in public comments is utility affordability.
An analysis from two universities in July found that the combination of data centers and cryptocurrency mining could raise electricity bills by 8% nationally, and as much as 25% in parts of Virginia. The report came from a collaboration between Carnegie Mellon University in Pennsylvania and North Carolina State University.
Virginia passed an electricity consumption surcharge specifically for data centers that began in July, in a budget compromise that preserved lucrative tax breaks for the data centers. Some localities with heavy concentrations are considering pauses, and at least one city has a moratorium.
This story was originally produced by Stateline, which is part of States Newsroom, a nonprofit news network which includes Wisconsin Examiner, and is supported by grants and a coalition of donors as a 501c(3) public charity.
The new school year has begun in the Empire State, and local school districts and bus companies have an additional five years to transition to all-electric fleets. But transportation leaders and policymakers say the clock is still ticking on major funding and infrastructure challenges that could make or break the ambitious mandate.
As part of the 2027 fiscal year budget passed in May, lawmakers extended the deadline for full compliance with New York’s electric school bus mandate from 2035 to 2040 and the requirement for all new school bus purchases to be zero-emission from 2027 to 2032. The move is widely viewed as a necessary reset for the nation’s largest school transportation system, but sources caution that time alone won’t solve the structural barriers districts face in electrifying their fleets.
“It’s great that both parts of the legislation were moved down five years, but the looming questions weren’t answered, and the school districts still have so many questions that the contractors really can’t answer,” said Corey Muirhead, executive vice president for New York City contractor Logan Bus. “They still don’t understand how the state funding is going to work and how they’re going to get reimbursed.”
A First Step, Not a Finish Line
For years, the state’s three primary school transportation associations—the New York Association for Pupil Transportation (NYAPT), New York School Bus Contractors Association (NYSBCA) and New York State Bus Distributors Association (NYSBDA)—have warned that the mandate, as originally written, was moving faster than the market, utilities and school budgets could reasonably support.
“Over the past five years, New York’s school transportation industry…has collaborated closely with Gov. [Kathy] Hochul’s administration and state lawmakers,” said lobbyist Andre Claridge, who represents NYAPT and the NYSBDA and formerly worked with the NYSBCA. “We have kept state leaders updated on our progress toward meeting the electric school bus mandate while highlighting the significant challenges faced by local school districts and industry partners.”
Claridge called the five-year extension “a welcome first step,” but stressed that the law’s language “leaves no room for innovation and requires revision.” He and other industry leaders are pushing for not only extended timelines, but increased state investment and regulatory flexibility that would allow other zero- and near-zero-emission technologies to count toward compliance.
“We opposed the mandate from the beginning and advocated for no mandate or a more realistic timeline. Finally, they heard us,” commented David Christopher, executive director of NYAPT. “We are now looking at how we will lobby going forward, because five years will go by quickly, and we need to provide realistic transition plans for the legislature to consider. Hopefully they will listen.”
Tim Flood, executive vice president for New York school bus contractor The Trans Group, said the compliance delay offers a little relief but noted the challenges “are still very real and numerous.”
“I think there will be changes to the mandate. I certainly hope there will be,” he continued, citing a belief that ongoing negotiations will add more delays between now and 2032. “But at this point and time, we have to move forward as it is exactly how it is.”
The new deadline could also be impacted by upcoming political election cycles, he added.
Meanwhile, the extension could provide breathing room to build out a more data-driven, district-by-district, company-by-company approach.
An electric Type A school bus operated by Trans Group company Educational Bus Transportation, Inc. on Long Island, New York.
Where Electric Buses Prove Viable Is a Diverse Playing Field
“There is no category of district based on what we’ve seen where an electric school bus is inherently not viable,” said Alok Disa, senior advisor for the New York State Energy and Research Development Authority’s (NYSERDA) clean transportation team.
NYSERDA is tasked by the state legislature with awarding funds to districts for purchasing electric school buses and related infrastructure as well as developing electrification plans for fleets. Disa pointed to real-world operations in rural, suburban and urban communities based on the organization’s experience work with school districts across the state.
About 50,000 school buses of all fuel types operate across New York. The split between large school buses and Type A “vans” is about 60-40.
Disa explained that more than 350 electric school buses are currently registered in New York state, up from roughly 250 to 275 the previous year, and 54 school districts currently have at least one electric school bus deployed. The World Resource Institute’s Electric School Bus Initiative says the number is more like 561 ESBs deployed in New York, with over 1,200 in the pipeline or “committed” to a school district or bus company through state or federal funding.
Despite the figures, most electric school buses operate outside New York City, countering the perception that electrification is only advancing in dense urban markets, Disa noted.
Early data from NYSERDA’s work with school districts on the fleet electrification planning program suggests that approximately 80 percent of routes statewide fall within the range today’s electric school buses can handle, once reasonable safety margins and cold-weather performance are factored in. In general, Disa said NYSERDA estimates that roughly 70 to 80 miles per day is the typical working range for many morning and afternoon home-to-school routes, even in cold regions such as the Adirondacks, Western New York and the Finger Lakes.
“We recommend that districts begin their electric school bus transition with some of their a.m./p.m. routes…rather than starting with extracurricular or longer trips,” he said, adding that districts operating buses in severe winter conditions have largely been able to keep those vehicles in service with appropriate planning and protocols.
Challenging Case in Point
In addition to being the transportation director for Naples Central School District, located about an hour drive southeast of Rochester, Pat Elwell also owns personal electric vehicles. That led him to push the district to apply for the first round of EPA Clean School Bus Program funding in 2022, around the same time the state legislature passed the electric mandate.
Elwell is also a bus mechanic, so his attention was finely tuned to how the ESBs would operate once the first two were delivered in March 2024. The district also has an electric Chevrolet Equinox used for students served by the McKinney-Vento Homeless Assistance Act.
“Our first issue came right out of the gate when our Level 2 chargers were taking 17-plus hours to charge our buses,” he shared with STN. “Working with the charger company, and getting a crash course in charger programing, we were able to fix this problem to where they [now] charge up within six to eight hours.”
Otherwise, he said maintenance has been “negligible,” though brake inspection intervals had to be cut in half to six months or 6,000 miles.
“I received a frantic radio transmission one morning from the driver that she smelled a burning smell on the bus. When the bus made it to the transportation facility, I quickly found that the smell was from the right front brake,” recounted Elwell, who is also the president of NYAPT’s Wayne-Finger Lakes chapter. “Turns out that with one pedal drive the drivers do not use the brakes much, therefore the pads had seized in the caliper bracket.”
While he said the ESBs only need one daily charge to operate all in-district and the one out-of-district route in 40- and 50-degree spring temperatures, an obvious challenge has been heating the ESBs in winter. Elwell said he learned the 10kWh heat pumps originally installed were too small for the cabin heating system. Even then, issues have remained.
“Working with the engineers at IC Bus they came to the conclusion that a 20-kWh pump should fix this problem,” he explained. “We did swap them out, and the difference in heat and battery consumption was significantly better. Fast forward to winter 2026, with sustained temps in the single digits and below zero, and the buses could not produce heat to be able to operate the buses safely and comfortably. Therefore, we moved the runs to diesel buses.”
He added that a repeated test of winter-time range in January and February 2025 documented the need for mid-day charging to complete back-to-back runs within the district service area, “which still holds true to this day,” he said.
The longest Naples routes are 80 miles round trip per day, Elwell continued, with the average morning and afternoon routes being 29 miles. In addition to the two EVs, the fleet currently consists of 13 diesel Type C school buses, six gasoline Type A vans, three SUVs, and a sedan.
“We currently use diesel/gasoline buses for out-of-district routes and field/athletic trips as the ESBs do not have the range,” he added.
So far, he said he has calculated a cost of around $0.57 per mile/kWh for his ESBs and $0.65 per mile for diesel. He added that he has been working with NYSERDA to develop a better formula.
Meanwhile, there’s the issue of support. It was only a couple of years ago Lion Electric filed for bankruptcy protection and was then acquired by Montreal investors, with the caveat that the electric school bus line would only be sold and serviced in Quebec. Other players in the segment have also fallen by the wayside, which creates concern for long-term customer support and warranty coverage.
“Some of those chassis, those companies aren’t even in business anymore, and I think that’s going to be one of the challenges,” commented Trans Group’s Flood. “Maybe not as much in the vehicles but in charging stations, in charging software. There are a lot of options, but how many are going to be there three, four or five years from now?”
“We don’t find a problem with electric buses because we are lucky enough to be in a location with sufficient power and have routes that are relatively short … [W]e’ve never had a bus where the batteries have been drained fully,” commented CEO Matt Berlin. “We are not worried about someone being stranded in the middle of rural New York. In the worst case, they’re in the middle of Queens.”
Berlin added that if NYCSBUS was running 100 ESBs, it might be a different story.
New York City formed the non-profit NYCSBUS in 2021 after acquiring the assets of school bus contractor Reliant Transportation. NYCSBUS operates a fleet of over 900 school buses with 888 routes. Twenty-nine of the school buses are electric. All so far are Type A’s consisting of 22 Micro Birds, six Type A Collins Bus (five with Lightning eMotors electric drives and one with a Motiv), and a RIDE Achiever.
As of the new school year, Berlin said the average daily route any school bus travels a day is 65 miles, which is normally within range of an ESB. A quarter of the 888 daily routes are 30 miles or shorter.
“We are not putting our EVs on every single route. We’re lucky that the city is compact,” he explained, adding that ESBs aren’t currently used on longer routes from Queens to New Jersey. “We’re lucky that regenerative braking benefits urban driving more than rural.”
He said the 29 Type A ESBs cost $0.13 per mile to operate, compared to $0.45 per mile for gasoline ($4.10 per gallon and 9 mpg) and $1 per mile for diesel ($6 per gallon and 6 mpg).
Like elsewhere in the state, Berlin said winter temperatures impact the range. He added that a proprietary route planning tool calculates which routes are best served by an ESB or an internal-combustion bus. The set goal is to return the former to the depot with a 20-percent charge remaining each day.
He also is looking at the viability of adding repowered school buses with electric batteries.
“We think the promise of repowering is terrific, if it worked right, and we haven’t seen it work. But if it did, it would cut the [purchase] price by a third,” he added. “Not to mention create jobs in New York City and train another generation of mechanics in how to work with these vehicles.”
Billions Still Needed for Buses and Charging
While route viability is trending positive, at least in some areas, dollars remain a major sticking point.
“It’s a massive capital outlay to reach this goal for many folks,” observed Flood at Trans Group. “For us as contractors, for school districts, whether they’re operating their own [fleet] or contracting or a combination of both, allocating or acquiring the capital to do those things … it’s just a challenge.”
Logan Bus’s Muirhead noted that new purchase costs of ESBs, could range from 30-percent to as high as 70-percent more than an internal combustion engine bus, putting school districts between a rock and a hard place.
“A school transportation’s budget was $2 million, and now it goes to $5 million?” he asked. “The way that I understand state reimbursement is that the district has to fund that $5 million, and then at the end of the year they get reimbursed by the state. This [is] the biggest elephant in the room that’s not addressed.”
Lobbyist Claridge said the industry has consistently told state leaders that fully funding the transition, “covering the billions required for new buses, charging infrastructure and power grid capacity,” he shared, is essential if New York is to convert what is “arguably the largest land-based mass transit system in the nation.”
Without that, he warned, the mandate risks outpacing realistic district budgets.
Another issue is the additional parking space needed for ESBs based on charging needs. Berlin said the NYCSBUS yard would need to be 25-percent larger to park ESBs with traditional bollards and in-ground chargers.
“That is certainly an issue for us. It is an additional cost of these vehicles,” he said. “It requires more parking, which is more space, and you know in New York City that’s expensive.”
He added NYCBUS has considered overhead chargers, portable charger carts and even induction or wireless charging.
So far, NYSERDA administers $500 million from the state Bond Act through the New York School Bus Incentive Program, with another $100 million added in the state budget, for a total of $600 million earmarked specifically for electric school buses and associated infrastructure.
Disa said that when NYSERDA incentives are combined with State Education Department aid formulas, “for a lot of districts…inclusive of the charging costs, it can be the equivalent, or even it can be cheaper to acquire an electric school bus than a diesel or combustion bus right now,” with no additional local tax impact.
Still, those funds represent the early days of a far longer and more expensive transition. With roughly 50,000 school buses operating statewide, and only a fraction on the road being electric or in the pipeline, tens of thousands of vehicles will still need to be replaced by 2040. Claridge argued that even with the extension, “meeting these goals will ultimately require additional time, state investment and flexibility.”
The U.S. Environmental Protection Agency is expected to announce the return of the Clean School Bus Program by next month. Nearly $2.5 billion must still be awarded nationwide. How much will go to New York school districts and bus companies remains to be seen.
Three electric school buses operated by school bus contractor Logan Bus in New York City.
Utilities and Infrastructure: ‘Early and Often’ Engagement
Beyond fleet costs, infrastructure and utility partnerships may pose the steepest near-term challenges for school districts and contractors.
“One of the critical elements that distinguish an electric school bus deployment compared to a traditional combustion bus is you do have to have a couple new stakeholders in the room, and one of those is your electric utility,” Disa said. “We recommend until we’re blue in the face that you’ve got to start talking to your electric utility early and often.”
Through NYSERDA-funded fleet electrification plans, school districts like Naples CSD are working with engineering firms to assess route energy needs, depot capacity, interconnection timelines and likely utility bill impacts. Disa said many districts currently have enough available capacity to support at least a few buses while longer-term upgrades are planned and financed, but those efforts take time and coordination.
Infrastructure upgrades are Naples CSD’s “biggest hurdle,” Elwell added. He said the district currently has adequate electricity for adding one more ESB to the fleet before utility provider NYSEG will need to upgrade the transmission line.
Claridge noted that utilities and the broader grid are central to whether the mandate can be implemented on the ground. “Fully transitioning [to electric buses] is impossible without adequate power grid capacity, charging infrastructure and funding,” he said.
Disa again pointed to NYSERDA’s fleet electrification plans led by consultants to help.
“These plans receive input from utilities, which can help determine infrastructure needs and next steps,” he explained.
For example, First Student announced Tuesday it is partnering with Con Edison and NYSERDA to demonstrate First Charge, the school bus contractor’s above-ground charging infrastructure. The organizations claim that installation could reduce infrastructure costs by 20 percent through less construction.
First Student is adding 40 ESBs in New York and is considering expanded solar and creating a smart energy hub.
“Deploying electric school buses at scale requires more than vehicles. It requires the right infrastructure, technology and utility partnership,” said Kevin Matthews, head of electrification at First Student. “Working with Con Edison, we’ve demonstrated a more efficient approach that can help overcome some of the cost and construction challenges that have traditionally made fleet electrification difficult in dense urban environments.”
Con Ed and NYSERDA are funding the project, which includes six bus vouchers through the New York School Bus Incentive Program and four retrofit buses.
Third-Party Testing and Policy Uncertainty
Another unresolved issue is a state law passed in last year’s budget requiring independent third-party testing of electric school bus range on specific routes and under varying conditions—a provision Claridge and other stakeholders describe as “overly complex” and disconnected from market realities.
Claridge said NYSERDA “continues to grapple with the implementation of this overly complex legislation,” and that the working group tasked with developing a regulatory framework “appears to have stalled” due to the law’s lack of clarity and practicality. He characterized the agency as “attempting to fit a square peg into a round hole.”
Disa said NYSERDA is using empirical fleet data and route planning results to answer the core questions lawmakers raised, such as how climate and terrain affect electric range. He noted NYSERDA convened a working group of industry stakeholders last September to help identify information that could benefit school bus operators and support industry stakeholders in finding their own paths to compliance.
‘Not a Time to Sit Back and Wait’
Everyone agrees that, despite the five-year extension, school districts cannot afford to wait.
Claridge emphasized that while the electric school bus mandate “has clear value,” the current policy framework “requires revision” to create space for innovation and realistic funding mechanisms.
Disa framed the extension as an opportunity for districts to move from planning to pilots and then to scale deployments, not as a pause button.
“Even though there is some extra time … it’s not a time to sit back and wait for years to elapse,” he said.
Districts that haven’t started a fleet plan should begin now, he added, while those with plans in hand should work with NYSERDA to get their first buses on the road and then scale up.
NYAPT’s Christopher said two important challenges remain.
“School districts vote on their budgets here, so regardless of the state mandates, if local district residents say no, buses are not purchased. This remains unresolved to date, putting a district in an untenable situation,” he explained. “Secondly, we have multiple utility companies operating in the state, and they all do business differently. It’s difficult for a school district or NYERSDA, for that matter, to sort through all the different rules and processes each utility follows in order to build out infrastructure.
“There needs to be more coordination and more attention to districts that need infrastructure built out if the mandate holds in five years,” he continued. “Districts need certainty from utilities regarding costs and timelines to secure voter approval and funding. It’s complicated to say the least.”
Fuel efficiency is deeply tied to the physical conditions that routes traverse. Not all miles are equal in those physical conditions. Road infrastructure plays a significant role in both school bus fuel consumption and vehicle wear.
In rural districts, infrastructure challenges can be even more pronounced, as noted by Rich Casto, transportation supervisor for Jackson County Schools in West Virginia. The school district uses data analytics to help mitigate associated challenges with sering an area of nearly 500 square miles along the Ohio River and Appalachian foothills.
Data analytics help districts evaluate alternative paths that may reduce wear, even if they require slight adjustments in routing.
Vehicle assignment is another emerging strategy. By matching higher-efficiency or newer buses to more demanding routes, districts can extend fleet life and reduce maintenance costs.
“When districts have this data at the route level, they can make smarter decisions about vehicle assignment [by] putting newer or higher-efficiency units on high-demand, high-wear routes and matching assets to routes in ways that extend useful life across the fleet,” said Craig Berndt, Geotab business segment manager for people transportation.
The connection between routing choices and maintenance outcomes is real and quantifiable.
“Districts that are actively using telematics to inform maintenance scheduling and route assignments are finding they can extend brake and tire replacement intervals meaningfully, and reduce unexpected breakdowns, which have their own significant cost in substitute vehicles and driver overtime,” he added.
Berndt pointed out that a 60-mile route on flat pavement is not the same maintenance proposition as a 60-mile route with six miles of gravel, three significant grades and a railroad crossing that requires a full stop every morning run.
Hard braking events, acceleration patterns and vehicle fault codes captured by GPS and engine diagnostic integration give maintenance teams early visibility into which routes are accelerating wear cycles, and fuel consumption varies significantly by road type, said Berndt.
“A bus working through frequent stop-and-go on a congested urban arterial will burn 20 to 30 percent more fuel per mile than the same bus on a steady rural highway run,” he said.
Kacy L. Davis, Sr., St. Louis (Missouri) Public Schools (SLPS) transportation director, noted not only do road conditions impact fuel use but maintenance cycles as well.
The Chenango Valley Central School District in Binghamton, New York, hones in on a few of its long out-of-district routes, keeps an eye on vehicles with excessive mileage and considers the impact of railroad crossings on bus routes.
“We have just a couple of long out-of-district routes,” noted Transportation Supervisor Jason Hibbard. “We focus on this a lot. We regularly switch buses on these routes so we can manage overall mileage on these vehicles.”
Last school year, the district saw an excessive increase in the usage of its SUVs for specialized transportation. It had to take the SUVs out of service as its two brand new vehicles are nearing the mileage of its 2017 vehicles.
“We are a 25-square mile district,” Hibbard explained. “Within that small space, we have 12 railroad crossings. Almost daily, we have to reroute based on slow or stopped trains.”
Kathy Calkins, North Clackamas (Oregon) School District transportation director, noted that consistent bus inspections help ascertain the impact of wear and tear from road infrastructure.
There is growing attention on the links between artificial intelligence and increased energy demands. But while the power-hungry data centers being built to support AI could potentially stress electricity grids, increase customer prices and service interruptions, and generally slow the transition to clean energy, the use of artificial intelligence can also help the energy transition.
For example, use of AI is reducing energy consumption and associated emissions in buildings, transportation, and industrial processes. In addition, AI is helping to optimize the design and siting of new wind and solar installations and energy storage facilities.
On electric power grids, using AI algorithms to control operations is helping to increase efficiency and reduce costs, integrate the growing share of renewables, and even predict when key equipment needs servicing to prevent failure and possible blackouts. AI can help grid planners schedule investments in generation, energy storage, and other infrastructure that will be needed in the future. AI is also helping researchers discover or design novel materials for nuclear reactors, batteries, and electrolyzers.
Researchers at MIT and elsewhere are actively investigating aspects of those and other opportunities for AI to support the clean energy transition. At its 2025 research conference, MITEI announced the Data Center Power Forum, a targeted research effort for MITEI member companies interested in addressing the challenges of data center power demand.
Controlling real-time operations
Customers generally rely on receiving a continuous supply of electricity, and grid operators get help from AI to make that happen — while optimizing the storage and distribution of energy from renewable sources at the same time.
But with more installation of solar and wind farms — both of which provide power in smaller amounts, and intermittently — and the growing threat of weather events and cyberattacks, ensuring reliability is getting more complicated. “That’s exactly where AI can come into the picture,” explains Anuradha Annaswamy, a senior research scientist in MIT’s Department of Mechanical Engineering and director of MIT’s Active-Adaptive Control Laboratory. “Essentially, you need to introduce a whole information infrastructure to supplement and complement the physical infrastructure.”
The electricity grid is a complex system that requires meticulous control on time scales ranging from decades all the way down to microseconds. The challenge can be traced to the basic laws of power physics: electricity supply must equal electricity demand at every instant, or generation can be interrupted. In past decades, grid operators generally assumed that generation was fixed — they could count on how much electricity each large power plant would produce — while demand varied over time in a fairly predictable way. As a result, operators could commission specific power plants to run as needed to meet demand the next day. If some outages occurred, specially designated units would start up as needed to make up the shortfall.
Today and in the future, that matching of supply and demand must still happen, even as the number of small, intermittent sources of generation grows and weather disturbances and other threats to the grid increase. AI algorithms provide a means of achieving the complex management of information needed to forecast within just a few hours which plants should run while also ensuring that the frequency, voltage, and other characteristics of the incoming power are as required for the grid to operate properly.
Moreover, AI can make possible new ways of increasing supply or decreasing demand at times when supplies on the grid run short. As Annaswamy points out, the battery in your electric vehicle (EV), as well as the one charged up by solar panels or wind turbines, can — when needed — serve as a source of extra power to be fed into the grid. And given real-time price signals, EV owners can choose to shift charging from a time when demand is peaking and prices are high to a time when demand and therefore prices are both lower. In addition, new smart thermostats can be set to allow the indoor temperature to drop or rise — a range defined by the customer — when demand on the grid is peaking. And data centers themselves can be a source of demand flexibility: selected AI calculations could be delayed as needed to smooth out peaks in demand. Thus, AI can provide many opportunities to fine-tune both supply and demand as needed.
In addition, AI makes possible “predictive maintenance.” Any downtime is costly for the company and threatens shortages for the customers served. AI algorithms can collect key performance data during normal operation and, when readings veer off from that normal, the system can alert operators that something might be going wrong, giving them a chance to intervene. That capability prevents equipment failures, reduces the need for routine inspections, increases worker productivity, and extends the lifetime of key equipment.
Annaswamy stresses that“figuring out how to architect this new power grid with these AI components will require many different experts to come together.” She notes that electrical engineers, computer scientists, and energy economists “will have to rub shoulders with enlightened regulators and policymakers to make sure that this is not just an academic exercise, but will actually get implemented. All the different stakeholders have to learn from each other. And you need guarantees that nothing is going to fail. You can’t have blackouts.”
Using AI to help plan investments in infrastructure for the future
Grid companies constantly need to plan for expanding generation, transmission, storage, and more, and getting all the necessary infrastructure built and operating may take many years, in some cases more than a decade. So, they need to predict what infrastructure they’ll need to ensure reliability in the future. “It’s complicated because you have to forecast over a decade ahead of time what to build and where to build it,” says Deepjyoti Deka, a research scientist in MITEI.
One challenge with anticipating what will be needed is predicting how the future system will operate. “That’s becoming increasingly difficult,” says Deka, because more renewables are coming online and displacing traditional generators. In the past, operators could rely on “spinning reserves,” that is, generating capacity that’s not currently in use but could come online in a matter of minutes to meet any shortfall on the system. The presence of so many intermittent generators — wind and solar — means there’s now less stability and inertia built into the grid. Adding to the complication is that those intermittent generators can be built by various vendors, and grid planners may not have access to the physics-based equations that govern the operation of each piece of equipment at sufficiently fine time scales. “So, you probably don’t know exactly how it’s going to run,” says Deka.
And then there’s the weather. Determining the reliability of a proposed future energy system requires knowing what it’ll be up against in terms of weather. The future grid has to be reliable not only in everyday weather, but also during low-probability but high-risk events such as hurricanes, floods, and wildfires, all of which are becoming more and more frequent, notes Deka. AI can help by predicting such events and even tracking changes in weather patterns due to climate change.
Deka points out another, less-obvious benefit of the speed of AI analysis. Any infrastructure development plan must be reviewed and approved, often by several regulatory and other bodies. Traditionally, an applicant would develop a plan, analyze its impacts, and submit the plan to one set of reviewers. After making any requested changes and repeating the analysis, the applicant would resubmit a revised version to the reviewers to see if the new version was acceptable. AI tools can speed up the required analysis so the process moves along more quickly. Planners can even reduce the number of times a proposal is rejected by using large language models to search regulatory publications and summarize what’s important for a proposed infrastructure installation.
Harnessing AI to discover and exploit advanced materials needed for the energy transition
“Use of AI for materials development is booming right now,” says Ju Li, MIT’s Carl Richard Soderberg Professor of Power Engineering. He notes two main directions.
First, AI makes possible faster physics-based simulations at the atomic scale. The result is a better atomic-level understanding of how composition, processing, structure, and chemical reactivity relate to the performance of materials. That understanding provides design rules to help guide the development and discovery of novel materials for energy generation, storage, and conversion needed for a sustainable future energy system.
And second, AI can help guide experiments in real time as they take place in the lab. Li explains: “AI assists us in choosing the best experiment to do based on our previous experiments and — based on literature searches — makes hypotheses and suggests new experiments.”
He describes what happens in his own lab. Human scientists interact with a large language model, which then makes suggestions about what specific experiments to do next. The human researcher accepts or modifies the suggestion, and a robotic arm responds by setting up and performing the next step in the experimental sequence, synthesizing the material, testing the performance, and taking images of samples when appropriate. Based on a mix of literature knowledge, human intuition, and previous experimental results, AI thus coordinates active learning that balances the goals of reducing uncertainty with improving performance. And, as Li points out, “AI has read many more books and papers than any human can, and is thus naturally more interdisciplinary.”
The outcome, says Li, is both better design of experiments and speeding up the “work flow.” Traditionally, the process of developing new materials has required synthesizing the precursors, making the material, testing its performance and characterizing the structure, making adjustments, and repeating the same series of steps. AI guidance speeds up that process, “helping us to design critical, cheap experiments that can give us the maximum amount of information feedback,” says Li.
“Having this capability certainly will accelerate material discovery, and this may be the thing that can really help us in the clean energy transition,” he concludes. “AI [has the potential to] lubricate the material-discovery and optimization process, perhaps shortening it from decades, as in the past, to just a few years.”
MITEI’s contributions
At MIT, researchers are working on various aspects of the opportunities described above. In projects supported by MITEI, teams are using AI to better model and predict disruptions in plasma flows inside fusion reactors — a necessity in achieving practical fusion power generation. Other MITEI-supported teams are using AI-powered tools to interpret regulations, climate data, and infrastructure maps in order to achieve faster, more adaptive electric grid planning. AI-guided development of advanced materials continues, with one MITEI project using AI to optimize solar cells and thermoelectric materials.
Other MITEI researchers are developing robots that can learn maintenance tasks based on human feedback, including physical intervention and verbal instructions. The goal is to reduce costs, improve safety, and accelerate the deployment of the renewable energy infrastructure. And MITEI-funded work continues on ways to reduce the energy demand of data centers, from designing more efficient computer chips and computing algorithms to rethinking the architectural design of the buildings, for example, to increase airflow so as to reduce the need for air conditioning.
In addition to providing leadership and funding for many research projects, MITEI acts as a convenor, bringing together interested parties to consider common problems and potential solutions. In May 2025, MITEI’s annual spring symposium — titled “AI and energy: Peril and promise” — brought together AI and energy experts from across academia, industry, government, and nonprofit organizations to explore AI as both a problem and a potential solution for the clean energy transition. At the close of the symposium, William H. Green, director of MITEI and Hoyt C. Hottel Professor in the MIT Department of Chemical Engineering, noted, “The challenge of meeting data center energy demand and of unlocking the potential benefits of AI to the energy transition is now a research priority for MITEI.”
Login
