If a data center is moving in next door, you probably live in the United States. More than half of all upcoming global data centers—as indicated by land purchased for data centers not yet announced, those under construction, and those whose plans are public—will be developed in the United States.

And these figures are likely underselling the near-term data-center dominance of the United States. Power usage varies widely among data centers, depending on land availability and whether the facility will provide xhttps://spectrum.ieee.org/data-center-liquid-cooling or mixed-use services, says Tom Wilson, who studies energy systems at the Electric Power Research Institute. Because of these factors, “data centers in the U.S. are much larger on average than data centers in other countries,” he says.

Wilson adds that the dataset you see here—which comes from the analysis firm Data Center Map—may undercount new Chinese data centers because they are often not announced publicly. Chinese data-center plans are “just not in the repository of information used to collect data on other parts of the world,” he says. If information about China were up-to-date, he would still expect to see “the U.S. ahead, China somewhat behind, and then the rest of the world trailing.”

One thing that worries Wilson is whether the U.S. power grid can meet the rising energy demands of these data centers. “We’ve had flat demand for basically two decades, and now we want to grow. It’s a big system to grow,” he notes.

He thinks the best solution is asking data centers to be more flexible in their power use, maybe by scheduling complex computation for off-peak times or maintaining on-site batteries, removing part of the burden from the power grid. Whether such measures will be enough to keep up with demand remains an open question.

This giant bubble on the island of Sardinia holds 2,000 tonnes of carbon dioxide. But the gas wasn’t captured from factory emissions, nor was it pulled from the air. It came from a gas supplier, and it lives permanently inside the dome’s system to serve an eco-friendly purpose: to store large amounts of excess renewable energy until it’s needed.

Developed by the Milan-based company Energy Dome, the bubble and its surrounding machinery demonstrate a first-of-its-kind “CO2 Battery,” as the company calls it. The facility compresses and expands CO₂ daily in its closed system, turning a turbine that generates 200 megawatt-hours of electricity, or 20 MW over 10 hours. And in 2026, replicas of this plant will start popping up across the globe.

We mean that literally. It takes just half a day to inflate the bubble. The rest of the facility takes less than two years to build and can be done just about anywhere there’s 5 hectares of flat land.

The first to build one outside of Sardinia will be one of India’s largest power companies, NTPC Limited. The company expects to complete its CO2 Battery sometime in 2026 at the Kudgi power plant in Karnataka, in India. In Wisconsin, meanwhile, the public utility Alliant Energy received the all clear from authorities to begin construction of one in 2026 to supply power to 18,000 homes.

And Google likes the concept so much that it plans to rapidly deploy the facilities in all of its key data-center locations in Europe, the United States, and the Asia-Pacific region. The idea is to provide electricity-guzzling data centers with round-the-clock clean energy, even when the sun isn’t shining or the wind isn’t blowing. The partnership with Energy Dome, announced in July, marked Google’s first investment in long-duration energy storage.

“We’ve been scanning the globe seeking different solutions,” says Ainhoa Anda, Google’s senior lead for energy strategy, in Paris. The challenge the tech giant has encountered is not only finding a long-duration storage option, but also one that works with the unique specs of every region. “So standardization is really important, and this is one of the aspects that we really like” about Energy Dome, she says. “They can really plug and play this.”

Google will prioritize placing the Energy Dome facilities where they’ll have the most impact on decarbonization and grid reliability, and where there’s a lot of renewable energy to store, Anda says. The facilities can be placed adjacent to Google’s data centers or elsewhere within the same grid. The companies did not disclose the terms of the deal.

Anda says Google expects to help the technology “reach a massive commercial stage.”

Getting creative with long-duration energy storage

All this excitement is based on Energy Dome’s one full-size, grid-connected plant in Ottana, Sardinia, which was completed in July. It was built to help solve one of the energy transition’s biggest challenges: the need for grid-scale storage that can provide power for more than 8 hours at a time. Called long-duration energy storage, or LDES in industry parlance, the concept is the key to maximizing the value of renewable energy.

When sun and wind are abundant, solar and wind farms tend to produce more electricity than a grid needs. So storing the excess for use when these resources are scarce just makes sense. LDES also makes the grid more reliable by providing backup and supplementary power.

The problem is that even the best new grid-scale storage systems on the market—mainly lithium-ion batteries—provide only about 4 to 8 hours of storage. That’s not long enough to power through a whole night, or multiple cloudy and windless days, or the hottest week of the year, when energy demand hits its peak.

After the CO2 leaves the dome, it is compressed, cooled, reduced to a liquid, and stored in pressure vessels. To release the energy, the process reverses: The liquid is evaporated, heated, expanded, and then fed through a turbine that generates electricity. Luigi Avantaggiato

Lithium-ion battery systems could be increased in size to store more and last longer, but systems of that size usually aren’t economically viable. Other grid-scale battery chemistries and approaches are in development, such as sodium-based, iron-air, and vanadium redox flow batteries. But the energy density, costs, degradation, and funding complications have challenged the developers of those alternatives.

Researchers have also experimented with storing energy by compressing air, heating up blocks or sand, using hydrogen or methanol, pressurizing water deep underground, and even dangling heavy objects in the air and dropping them. (The creativity devoted to LDES is impressive.) But geologic constraints, economic viability, efficiency, and scalability have hindered the commercialization of these strategies.

The tried-and-true grid-scale storage option—pumped hydro, in which water is pumped between reservoirs at different elevations—lasts for decades and can store thousands of megawatts for days. But these systems require specific topography, a lot of land, and can take up to a decade to build.

CO2 Batteries check a lot of boxes that other approaches don’t. They don’t need special topography like pumped-hydro reservoirs do. They don’t need critical minerals like electrochemical and other batteries do. They use components for which supply chains already exist. Their expected lifetime stretches nearly three times as long as lithium-ion batteries. And adding size and storage capacity to them significantly decreases cost per kilowatt-hour. Energy Dome expects its LDES solution to be 30 percent cheaper than lithium-ion.

China has taken note. China Huadian Corp. and Dongfang Electric Corp. are reportedly building a CO₂-based energy-storage facility in the Xinjiang region of northwest China. Media reports show renderings of domes but give widely varying storage capacities—including 100 MW and 1,000 MW. The Chinese companies did not respond to IEEE Spectrum’s requests for information.

“What I can say is that they are developing something very, very similar [to Energy Dome’s CO2 Battery] but quite large in scale,” says Claudio Spadacini, Energy Dome’s founder and CEO. The Chinese companies “are good, they are super fast, and they have a lot of money,” he says.

Why is Google investing in CO2 Batteries?

When I visited Energy Dome’s Sardinia facility in October, the CO₂ had just been pumped out of the dome, so I was able to peek inside. It was massive, monochromatic, and pretty much empty. The inner membrane, which had been holding the uncompressed CO₂, had collapsed across the entire floor. A few pockets of the gas remained, making the off-white sheet billow up in spots.

Meanwhile, the translucent outer dome allowed some daylight to pass through, creating a creamy glow that enveloped the vast space. With no structural framing, the only thing keeping the dome upright was the small difference in pressure between the inside and outside air.

“This is incredible,” I said to my guide, Mario Torchio, Energy Dome’s global marketing and communications director.

“It is. But it’s physics,” he said.

Outside the dome, a series of machines connected by undulating pipes moves the CO₂ out of the dome for compressing and condensing. First, a compressor pressurizes the gas from 1 bar (100,000 pascals) to about 55 bar (5,500,000 pa). Next, a thermal-energy-storage system cools the CO₂ to an ambient temperature. Then a condenser reduces it into a liquid that is stored in a few dozen pressure vessels, each about the size of a school bus. The whole process takes about 10 hours, and at the end of it, the battery is considered charged.

To discharge the battery, the process reverses. The liquid CO₂ is evaporated and heated. It then enters a gas-expander turbine, which is like a medium-pressure steam turbine. This drives a synchronous generator, which converts mechanical energy into electrical energy for the grid. After that, the gas is exhausted at ambient pressure back into the dome, filling it up to await the next charging phase.

Energy Dome engineers inspect the dryer system, which keeps the gaseous CO₂ in the dome at optimal dryness levels at all times.Luigi Avantaggiato

It’s not rocket science. Still, someone had to be the first to put it together and figure out how to do it cost-effectively, which Spadacini says his company has accomplished and patented. “How we seal the turbo machinery, how we store the heat in the thermal-energy storage, how we store the heat after condensing…can really cut costs and increase the efficiency,” he says.

The company uses pure, purpose-made CO₂ instead of sourcing it from emissions or the air, because those sources come with impurities and moisture that degrade the steel in the machinery.

What happens if the dome is punctured?

On the downside, Energy Dome’s facility takes up about twice as much land as a comparable capacity lithium-ion battery would. And the domes themselves, which are about the height of a sports stadium at their apex, and longer, might stand out on a landscape and draw some NIMBY pushback.

And what if a tornado comes? Spadacini says the dome can withstand wind up to 160 kilometers per hour. If Energy Dome can get half a day’s warning of severe weather, the company can just compress and store the CO₂ in the tanks and then deflate the outer dome, he says.

If the worst happens and the dome is punctured, 2,000 tonnes of CO₂ will enter the atmosphere. That’s equivalent to the emissions of about 15 round-trip flights between New York and London on a Boeing 777. “It’s negligible compared to the emissions of a coal plant,” Spadacini says. People will also need to stay back 70 meters or more until the air clears, he says.

Worth the risk? The companies lining up to build these systems seem to think so.

This article appears in the January 2026 print issue as “Grid-Scale CO₂ Batteries Will Take Off in 2026.”

The United States aims to embark on its most active new nuclear construction program since the 1970s. In its most high-dollar nuclear deal yet, the Trump administration in October launched a partnership to build at least US $80 billion worth of new, large-scale nuclear reactors, and chose Westinghouse Electric Company and its co-owners, Brookfield Asset Management and Cameco, for the job.

The money will support the construction of AP1000s, a type of pressurized water reactor developed by Westinghouse that can generate about 1,110 megawatts of electric power. These are the same reactors as units 3 and 4 at the Vogtle nuclear plant in Georgia, which wrapped up seven years behind schedule in 2023 and 2024 and cost more than twice as much as expected—about $35 billion for the pair. Along the way, Westinghouse, based in Cranberry Township, Penn., filed for Chapter 11 bankruptcy protection.

Chief executives of investor-owned utilities know that if they were to propose committing to similar projects on the same commercial terms, they’d be sacked on the spot. As a result, the private sector in the United States has been unwilling to take on the financial risk inherent in building new reactors.

The $80 billion deal with the federal government represents the U.S. nuclear industry’s best opportunity in a generation for a large-scale construction program. But ambition doesn’t guarantee successful execution. The delays and cost overruns that dogged the Vogtle project present real threats for the next wave of reactors.

Streamlining AP1000 Reactor Construction

What might be different about the next set of AP1000s? On the positive side, delivering multiple copies of the same reactor ought to create the conditions for a steady decline in costs. Vogtle Unit 3 was the first AP1000 to be built in the United States, and the lessons learned from it resulted in Vogtle Unit 4 costing 30 percent less than Unit 3. (Six AP1000s are currently operating outside the United States, and 14 more are under construction, according to Westinghouse.)

There’s been a bipartisan effort in the United States to streamline regulatory procedures to ensure that future projects won’t be delayed by the same issues that hampered Vogtle. The Accelerating Deployment of Versatile, Advanced Nuclear for Clean Energy (ADVANCE) Act that was signed into law by former U.S. President Joe Biden in 2024 includes several measures intended to improve processes at the Nuclear Regulatory Commission (NRC).

The last nuclear reactors to be built in the United States—Vogtle Units 3 and 4 in Waynesboro, Georgia—were completed seven years behind schedule and cost more than twice as much as expected.Georgia Power Co.

That included a mandated change in the NRC’s mission statement, setting a goal of “enabling the safe and secure use and deployment of civilian nuclear energy technologies.” It was a symbol of Congress’s intent to encourage the commission to support nuclear development.

In May, President Trump built on that legislation with four executive orders intended to speed up reactor licensing and accelerate nuclear development—a framework that has yet to be tested in practice. In November, the NRC published regulations setting out how it planned to implement the president’s orders. The changes are focused on removing redundant and duplicative rules.

One of Trump’s orders included a series of provisions intended to help build the U.S. nuclear workforce, but it’s clear that will be a challenge. The momentum gained in training skilled workers during the construction at Vogtle is already dissipating. Without other active new reactor projects to move on to immediately in the United States, many of the people who worked there have likely gone into other sectors, such as liquified natural gas (LNG) plants.

Around the time that construction was wrapping up at Vogtle, many employers in the industry were already reporting difficulties in finding the staff they need, according to the Department of Energy’s 2025 United States Energy and Employment Report. Surveyed in 2024, 22 percent of employers in nuclear construction said it was “very difficult” to hire the workers they needed, and 63 percent said it was “somewhat difficult.” In nuclear manufacturing, 63 percent of employers said hiring was “very difficult.”

If reactor construction really begins to pick up, there is clearly a danger that those numbers will rise.

U.S. Nuclear Power Expansion Plans

So just how many reactors will $80 billion buy? Assuming an average of $16 billion per AP1000—slightly less than for Vogtle, and allowing for cost reductions from economies of scale and learning-by-doing—the plan would mean five new reactors. That would represent an increase of about 5.7 percent in total U.S. nuclear energy generation capacity, if all the reactors currently in service remain online.

The full details of the $80 billion deal, including the precise allocation of financing and risk-sharing, have not been specified. But Westinghouse’s co-owner, Brookfield, did disclose that the partnership includes profit-sharing mechanisms that will give the U.S. government some of the upside if the initiative succeeds.

The Washington Post reported that after the U.S. signs the final contracts for $80 billion worth of new reactors, it will be entitled to 20 percent of all of Westinghouse’s returns over $17.5 billion. And if Westinghouse’s valuation surpasses $30 billion, the administration can require it to be floated on the stock market. If that happens, the government will get a 20 percent stake.

Enriched uranium is loaded at Vogtle Unit 4.Georgia Power Co.

Japan’s government is also playing a key role. As part of a $550 billion U.S.-Japan trade deal struck in July, the Japanese government pledged large-scale investment in U.S. energy, including nuclear. Japanese companies, including Mitsubishi Heavy Industries, Toshiba Group, and IHI Corp., are interested in investing up to $100 billion in the United States to support the construction of new AP1000s and small modular reactors (SMRs), the two governments said.

The Westinghouse deal supports a range of the administration’s objectives, including power for AI and investment and job creation in the American industrial sector. The focus on AP1000s also makes it possible to rely on U.S.-produced fuel, strengthening energy security. (Many of the designs for SMRs, which have garnered a considerable amount of excitement globally, use high-assay, low-enriched uranium (HALEU) fuel, which is not currently produced on a large scale in the United States.)

U.S. Nuclear Energy Investment

There have been other recent moves to add additional nuclear capacity in the United States. Santee Cooper, a South Carolina utility, announced plans for completing the construction of two AP1000 reactors that had been abandoned in 2017 at the V.C. Summer site in Jenkinsville, S.C.

Separately, Google announced in October a deal with NextEra Energy to reopen a 615-MW nuclear plant in Iowa. The Duane Arnold Energy Center was shut down in 2020, and the aim is to have it operational again by the first quarter of 2029. Google has agreed to buy a share of the plant’s output for 25 years.

Construction of two AP1000 reactors at the V.C. Summer nuclear site in Jenkinsville, S.C., was abandoned in 2017 after delays and cost overruns. Executives leading the projects were charged with fraud. Chuck Burton/AP

But the plans that have been announced so far pale in comparison to the Trump administration’s nuclear ambitions. Earlier this year, Trump set a goal of adding a whopping 300 gigawatts of nuclear capacity by 2050, up from a little under 100 GW today. That would mean much stronger growth than is currently projected in Wood Mackenzie’s forecasts, which show a near-doubling of U.S. nuclear generation capacity to about 190 GW in 2050.

The main driver behind the Trump administration’s interest in nuclear is its ambitions for artificial intelligence. Chris Wright, the U.S. energy secretary, has described the race to develop advanced AI as the Manhattan Project of our times, critical to national security and dependent upon a steep increase in electricity generation. Speaking to the Council on Foreign Relations in September, Wright promised: “We’re doing everything we can to make it easy to build power generation and data centers in our country.”

One of the hallmarks of the Trump administration has been its readiness to intervene in markets to pursue its policy goals. Its nuclear strategy exemplifies that approach. In many ways, the Trump administration is acting like an energy company: using its financial strength and its convening power to put together a deal that covers the entire nuclear value chain.

Throughout the history of nuclear power, the industry has worked closely with governments. But the federal government effectively taking a commercial position in the development of new reactors would be a first for the United States. In the first wave of U.S. reactor construction in the 1970s, federal government support was limited to R&D, uranium mining and enrichment, and indemnifying operators against the risk of nuclear accidents.

Before the partial deregulation of U.S. electricity markets that began in the 1990s, utilities could develop nuclear plants with the assurance that the costs could be recovered from customers, even if they went far over budget. With many key markets now at least partially deregulated, nuclear project developers will need other types of guarantees to secure financing and move forward.

The first new plants that result from the $80 billion deal will come online years after Trump has left office. But they could play an important role in boosting U.S. electricity supply and developing advanced AI for decades.

In 1950, the English mathematician Alan Turing devised what he called “the imitation game.” Later dubbed the Turing test, the experiment asks a human participant to conduct a conversation with an unknown partner and try to determine if it’s a computer or a person on the other end of the line. If the person can’t figure it out, the machine passes the Turing test.

Power grid operators are now preparing for their own version of the game. Virtual power plants, which concatenate small, distributed energy resources, are increasingly being tapped to balance electricity supply and demand. The question is: Can they do their job as well as conventional power plants?

Grid operators can now find out by running these power plants through a Turing-like test called the Huels. To pass the Huels test, the performance of a virtual power plant must be indistinguishable from that of a conventional power plant. A human grid operator serves as the judge.

Virtual power plant developer EnergyHub, based in Brooklyn, N.Y., developed the test and outlined it in a white paper released today. “What we’re really trying to do is fool the operators into feeling that these virtual power plants can act and feel and smell like conventional power plants,” says Paul Hines, chief scientist at EnergyHub. “This is a kind of first litmus test.”

What Are Virtual Power Plants (VPPs)?

The virtual-versus-conventional power plant question is a timely one. Virtual power plants, or VPPs, are networks of devices such as rooftop solar panels, home batteries, and smart thermostats that come together through software to collectively supply or conserve electricity.

Unlike conventional power generation systems, which might crank up one big gas plant when electricity demand peaks, VPPs tap into small, widely disbursed equipment. For example, a VPP might harness electricity from hundreds of plugged-in electric vehicles or rooftop solar panels. Or it might direct smart thermostats in homes or businesses to turn down heat or cooling systems to reduce demand.

The technology is emerging at a time when concerns over data centers’ electricity demand is hitting a fever pitch. The consultancy BloombergNEF estimates data-center energy demand in the United States will reach 106 gigawatts by 2035–a 36 percent jump from what it had projected just seven months ago.

How utilities and grid operators will meet the growing demand is unclear and faces challenges on many fronts. Turbines for natural gas plants are backordered, and new nuclear reactors are still years away. Wind and solar, while cheap and fast to build, don’t produce the 24/7 electricity that data centers demand and face an uphill political battle under the Trump administration.

All of this together has created an opening for VPPs, which could add gigawatts to the grid without significantly jacking up electricity rates. “It’s a political issue. If you said you’re going to get electricity costs under control, this is literally the only way to do it in 12 months,” says Jigar Shah, a clean energy investor at Multiplier in Washington, D.C., who led the U.S. Department of Energy’s Loan Programs Office under the Biden administration.

VPPs could also reduce utilities’ need to invest in distribution equipment, avoiding supply chain shortages and inflated costs, Shah says. “There is no other idea that you could possibly deploy in 12 months that would have that big of an impact,” he says.

According to a 2024 U.S. Department of Energy report, VPPs could provide between 80 and 160 gigawatts of capacity across the U.S. by 2030—enough to meet between 10 and 20 percent of peak grid demand.

How Can VPPs Gain Grid Operator Trust?

But first, VPP developers have to win over grid developers. Benchmarks like the Huels test are crucial to building that trust. “In order for us to build our reliance on VPPs, they do need to pass the Huels test, and operators need to be able to count on” the VPPs delivering power when called upon, said Lauren Shwisberg, a principal in the nonprofit research group Rocky Mountain Institute who co-authored a recent report on VPPs and was not involved in the development of the test.

Matthias Huels, an engineer who spent more than four years at EnergyHub, first came up with the idea for the test in 2024. After workshopping the idea with colleagues and, somewhat ironically, ChatGPT, Huels presented the concept to the company.

Huels designed the test subjectively. Currently, in its earliest iteration, it appears to follow a guideline akin to the Supreme Court’s “I know it when I see it” test for what distinguishes pornography from erotic art. That is to say: Passing the test depends on who’s judging. If a grid operator finds the power from a VPP as dependable as electricity from an actual power plant burning gas to produce electrons, then the VPP has passed.

There are four levels to the Huels test. To reach level 1, a VPP must be able to shave off demand from the grid by, for example, successfully scheduling smart thermostats to dial down when the grid faces maximum demand. To reach level 2, a VPP must be able to respond to market and grid data and dial down demand when prices hit a certain level or tap into solar panels or batteries when power is needed. Human decision makers are involved at these levels.

Passing the Huels test comes at level 3. That’s when a VPP can function automatically because it’s proven reliable enough to be indistinguishable from a gas peaker plant–the type of power station that comes online as backup only when the grid is under stress. Passing level 4 involves VPPs acting fully autonomously to adjust output based on a number of actively changing variables throughout the day.

“The imitation game that Alan Turing came up with was: Can a computer fool an interrogator to think it’s actually human even though it’s a computer,” Hines says. “We propose this idea of a test that would allow us to say: Can we fool a grid operator into thinking that the thing that’s actually solving their problems is this aggregation of many devices instead of a big gas plant?”

Can VPPs Mimic Gas Peaker Plants?

Peaker plants only generate power about 5 percent of the time over their lifespans. That makes them easier for VPPs to mimic because, like peaker plants, the limited amount of power that can be made available by demand response or harvested from batteries only provides bursts of power that last a few hours at a time.

Far more difficult is stacking up to a full-scale gas plant, which operates 65 percent of the time or more, or a nuclear plant, which usually operates at least 95 percent of the time. Getting there would involve equipping a VPP network with long-duration storage that could be powered up during the day when solar panels are at peak output and discharged all night long. “You start talking about VPPs with large amounts of batteries that can run 365 days per year,” Hines says. “That’s a road we can go down.”

EnergyHub has been putting its VPP systems through the Huels test. Last year, EnergyHub successfully ran trials with Arizona Public Service, Duke Energy in North Carolina, and National Grid in Massachusetts. In Arizona, EnergyHub’s software dialed into homes with solar panels and smart thermostats and ran air conditioners to “pre-cool” houses during the day when the sun was generating lots of electricity. This allowed the state’s biggest utility to reduce demand during peak hours when residents would typically return home from work to turn on televisions and crank up their air conditioners.

“You have too much power in the middle of the day because of solar, then the early evening comes and you get people ramping up their evening loads right as the solar is ramping down,” Hines says. “You need something that can feather through that schedule. We created something that can do this.”

That lands the company somewhere between a 2 and 3 on the Huels testing scale. Passing level 3 “is going to take a few years,” Hines says.