Global warming, climate change, all these things are just a dream come true for politicians. I deal with evidence and not with frightening computer models because the seeker after truth does not put his faith in any consensus. The road to the truth is long and hard, but this is the road we must follow. People who describe the unprecedented comfort and ease of modern life as a climate disaster, in my opinion have no idea what a real problem is.
Across the country, voters are asking the same basic questions: Will the AI boom raise my electricity bill? Will it strain the grid during peak demand? Will my community be asked to absorb the land-use, noise, and infrastructure burden while someone else captures the upside?
Server racks in data center. Backup, hosting, mainframe, farm and computer rack concept. 3D rendering
When members of Congress return home for the summer, few issues will confront them more quickly than the explosion of data center development.
Across the country, voters are asking the same basic questions: Will the AI boom raise my electricity bill? Will it strain the grid during peak demand? Will my community be asked to absorb the land-use, noise, and infrastructure burden while someone else captures the upside?
Those are fair questions. They are not anti-technology or anti-growth. They are what people ask when they’re promised an AI future without an explanation of who pays for it.
Conservatives should take those concerns seriously. Families should not be forced to subsidize billion-dollar computing facilities. Local communities should not be kept in the dark. And lawmakers should not pretend that every objection to a data center is just NIMBYism. The backlash is real. In our June 2026 national survey, voters opposed an AI data center being built in their community by more than two to one — 62% to 27% — with nearly half strongly opposed.
Right now, solar and battery energy storage are among the only major energy resources that can be deployed quickly enough and affordably enough to keep pace with rising demand. In fact, they have accounted for more than 80% of all new power added to the grid during the Trump administration.
Voters are ready for that buildout. In the same survey in which they rejected data centers, nearly two-thirds of Americans said they would welcome a solar farm in their own community — support on par with an Amazon distribution center or a new manufacturing plant. Americans are not against building. They are against being handed the costs of the AI boom without the power to pay for it.
The scale of the challenge is not theoretical. A June 2026 analysis from Lawrence Berkeley National Laboratory found that data centers could more than double their electricity use by 2030 and represent more than 40% of U.S. electricity demand growth over the next five years. The question is no longer whether AI needs significant power. It is whether America will build it in time.
That is why conservatives should champion energy abundance. Natural gas matters. Nuclear matters. Transmission matters. So do solar and battery storage, because they are practical tools that can be deployed relatively quickly, add meaningful capacity, and help meet demand during the very hours when the grid is under pressure.
This is not a partisan issue. In our January survey of the likely 2026 electorate, solar was one of the best-liked industries in America — viewed favorably by 67% of voters. Trump voters view solar companies favorably by nearly two to one, and in rural America, solar is favorably viewed by a wide margin.
In fast-growing states, solar is already helping utilities and grid operators keep pace with rising load, while batteries are increasingly shifting that power into the evening hours, when families come home, air conditioners are running, and solar output begins to fall.
Texas offers a useful example. The U.S. Energy Information Administration (EIA) says wind and solar met 36% of ERCOT demand in the first nine months of 2025, while batteries increasingly helped supply evening electricity as solar output declined. EIA also forecasts that utility-scale solar in ERCOT will exceed coal generation on an annual basis for the first time in 2026.
Some critics will say that these projects simply create another land-use problem. That objection deserves a factual answer, too. A recent interactive tool released by the Solar Energy Industries Association (SEIA) does just that, showing that solar uses just 0.04% of total U.S. land and 0.07% of U.S. farmland, with no state using more than 0.5% of its prime farmland for solar. Not every project belongs everywhere, and communities deserve transparency and respect for legitimate local concerns. But conservatives should be wary of turning every concern into a veto over the infrastructure America needs to power its future.
The alternative to energy abundance isn’t standing still. It’s falling behind. It means delays, special deals, local moratoria, higher prices, and growing resentment. That is bad for families, bad for the grid, and bad for American competitiveness at a moment when China is not slowing down.
Lawmakers going home this summer do not need a better slogan on AI. They need a better answer on power. Large loads should pay their fair share. Communities should get straight answers. And America should build more energy of every kind, including the solar and storage projects already helping fast-growing states, many of which are traditional conservative strongholds, add real power on timelines that match the urgency of the moment.
AI won’t run on talking points. It will run on affordable, abundant electricity. The market is already pointing us toward the fastest, most affordable path to more power. Let’s take it.
You’ve heard the stories. Data centers are loud, water-guzzling monsters chewing up rural landscapes. They promise jobs but deliver none. They’ll suck the grid dry, jack up everyone’s electric bill, overload the power grid, and leave abandoned warehouses when the AI bubble bursts. Opponents paint them as the latest corporate scam, with Chinese Communist Party–backed accounts flooding social media to amplify every complaint from low information voters.
You’ve heard the stories. Data centers are loud, water-guzzling monsters chewing up rural landscapes. They promise jobs but deliver none. They’ll suck the grid dry, jack up everyone’s electric bill, overload the power grid, and leave abandoned warehouses when the AI bubble bursts. Opponents paint them as the latest corporate scam, with Chinese Communist Party–backed accounts flooding social media to amplify every complaint from low information voters.
Meanwhile, the Big Tech acolytes tell us these monstrosities will unlock human flourishing to a degree never before seen in history. The AI revolution will free humanity to create instead of having to punch a clock as a slave to the grind. It will transform industries, contribute trillions to the global economy, and even heal the planet by slowing global warming. All at a relatively low low strain on resources, unlike the claims to the contrary by those opposed to data centers in their back yard.
So who’s right?
The social media furor has gotten so loud that local governments are hesitating, even while the technology reshapes the economy.
The sheer volume of misinformation surrounding the issue has marred any sort of thoughtful debate about the pros and cons of data centers.
To be fair, this is not to advocate for one side of the debate or the other. The purpose of this exercise is to examine the BS and offer actual facts so the debate can proceed with true understanding, not hype and fearmongering.
This graphic made the rounds recently on social media. Regardless of verifying whether the claims hold any statistical significance, it indicates the sheer volume of misinformation and disinformation emanating from both sides of the data center divide:
Luckily, we have an energy expert who can address these myths and give us the facts we need to make informed decisions about where to site the inevitable expansion of data centers.
One thing is certain: the race for global AI dominance will continue, as will the need domestic solutions.
Why All the Myths?
Glenn Davis has spoken in Delaware, Maryland, Pennsylvania, West Virginia, and Virginia, cutting through the fog. A former Virginia Beach city councilman, state assemblyman, and director of the Virginia Department of Energy, he doesn’t defend data centers or attack them. He simply insists on facts over recycled talking points.
In an interview with Restoration News, Davis said, “A lot of the people that are opposed to data centers have never seen a data center. They’ve never lived near a data center.” He believes most of the hype in Virginia—the state with the most data centers—comes from a deep blue Democrat stronghold, Loudoun County.
Davis has years of experience in dismantling the hype. As he points out, we’ve had server farms and data centers for years—in fact, 70 percent of all global internet traffic runs through data centers in Virginia, getting their start with government contracts and at the headquarters for America Online (AOL).
“Start with the noise,” Davis said. “Older facilities from the early 2000s sit cheek-by-jowl with homes in Loudoun because the county never updated zoning in time. Those plants hum from servers and cooling systems built before more recent sound attenuation became standard.”
“I had a friend that did Bitcoin mining,” Davis recalled, “and it was pretty loud because they didn’t have the sound attenuation that they have today.” Modern centers are different. To prove this, Davis will take a West Virginia delegation to Loudoun County with decibel meters to prove the point. “The difference between yesterday’s tech and today’s is night and day. Most neighbors wouldn’t know a new facility was there unless someone pointed it out.”
The myth that every data center makes for a noisy neighbor comes straight from a handful of outdated Loudoun County sites that no one would build the same way today. The water myth follows the same pattern. Critics claim data centers guzzle precious water resources. But as Davis points out, the complaints stem from those same old early 2000s facilities that used open-loop systems.
Today’s centers rely on closed-loop cooling that recycles water and dramatically cuts consumption. The technology has moved on, but the talking points haven’t.
Then comes the tax panic. Critics claim data centers starve local budgets or shift burdens onto homeowners. The opposite is true. In Mecklenburg County, a single Microsoft facility delivered enough property-tax revenue to fund $134 million in immediate school renovations—middle and high schools first, elementary now. Davis noted that data centers now account for roughly 40 percent of property taxes in some Virginia localities.
Loudoun County itself enjoys one of the lowest property-tax rates in the state precisely because of revenue from data centers. “If people understood what their property taxes in Loudoun County would be if they didn’t have data centers, they’d be shocked,” he said flatly.
That broadened tax base doesn’t just patch potholes; it rebuilds entire school systems and funds community projects without the usual trade-offs. Rural counties that land even one or two facilities suddenly have the money to improve roads, parks, and public services that used to depend on raising taxes on existing homeowners. It’s the rare economic development win that pays its own way and then some.
The jobs argument gets twisted, too. Yes, direct employment matters—maintenance crews, construction booms when clusters form—but the real payoff is subtler. Data centers don’t import thousands of new families demanding new roads, schools, and traffic lights the way a factory or housing development does. They deliver massive annual tax revenue with almost zero additional service burden. That money then improves schools, builds amenities, and creates the very culture that attracts advanced manufacturing and the skilled workforce Virginia has struggled to keep.
The Source of the Myths
Kevin O’Leary—Mr. Wonderful from the reality TV show Shark Tank—has started sounding the alarm about foreign influence in the online debate about data centers. After a large Utah facility he was invested in was beset by delays and downsizing, he began investigating.
He pointed to a sudden surge in negative comments on social media platforms starting in early May 2026—“tens of thousands” of messages on Instagram and X—with patterns suggesting coordination: shared IP addresses, bot activity, and “nefarious accounts from out of the country.” He says the organizations pushing anti-data center propaganda have ties to Arabella Network and Neville Roy Singham, both of which act as proxies for the Chinese Communist Party.
O’Leary notes that China has a clear and vital interest in seeing the U.S. fall behind in data center development, even as President Trump has declared winning the AI race a vital national interest.
Watching Jobs Disappear—and a Possible Solution
Davis lived this reality on Virginia Beach city council. He watched manufacturers beg for better schools and downtown vibrancy before they’d relocate talent from Germany or elsewhere. “Data centers break that chain of chicken and egg because they’ll come to your area without you having to have a huge workforce there,” he explained. “They’ll dump an enormous amount of annual tax revenue into the locality that generally you use to invest in your schools, invest in the area, and now you’re going to bring in advanced manufacturing and other types of jobs.
Davis continued to tick off the economic benefits he’s watched develop:
Community colleges are already responding with new HVAC, electrician, and trades programs, retooling workers mid-career. In Hampton Roads, Navy retirees bring precision skills that fit perfectly. The clustering effect multiplies this: one center draws others, turning three-year construction cycles into sustained local work without overwhelming small-town infrastructure.
Virginia doesn’t have to wait years to build a manufacturing base; the data centers create the revenue stream that funds the training pipeline today. Localities can finally invest in the education and amenities that draw the very people manufacturers need.
The loudest myth of all is the power grab. Data centers will devour the grid, the story goes, and ratepayers will foot the bill while the tech giants laugh. Davis has negotiated these deals at the state Department of Energy level. The reality is the opposite. “The industry’s price elasticity is unlike anything else. Reliability matters more than cost,” he said. “That is why Microsoft is building a 1. 4-gigawatt facility in West Virginia entirely behind the meter on natural gas—no grid connection, no utility middleman. Virginia law already allows data centers to bring their own power.”
The most exciting development revolves around the revival of America’s nuclear power industry. West Virginia passed a data-center microgrid statute giving them retail choice. Even more striking, data centers are bankrolling the nuclear renaissance that politicians have talked about for decades. Small modular reactors (SMRs) and advanced designs like the AP1000 are moving forward because hyperscalers are writing billion-dollar checks upfront and signing long-term power purchase agreements. “The first set of these new technologies will be for data centers. It won’t be for you and I,” Davis explained.
The economics are sound, though. The high first-of-a-kind costs get absorbed by companies that can pass them on to their customers. Once the nth-of-a-kind price drops—roughly 20 percent per unit after the first few—the technology becomes viable for utilities and residential ratepayers. In the last PJM capacity auction, the regional grid was short 6.2 gigawatts; 5.1 gigawatts of that shortfall came solely from new data-center demand. Far from draining the system, they are forcing the build-out of reliable generation that eventually benefits everyone.
The data centers are paying to solve the very capacity crunch the rest of us have been complaining about for years. Virginia already routes 70 percent of global data-center traffic. That didn’t happen by accident; it began with early internet backbone points like MAE-East and the original AOL campus. The facilities of the future—gigawatt-scale campuses, some already announced in Canada—will be larger still.
The industry’s own capital expenditures fund the very power solutions the rest of the grid needs. A 5-gigawatt data center campus is on the drawing board right now. That’s one-fifth of Virginia’s entire current statewide energy usage, all in a single project. And the hyperscalers aren’t waiting around for someone else to solve the power problem—they’re writing the checks themselves.
This Isn’t a Bubble
Davis invokes Jevons paradox: efficiency breeds more consumption. The first computers filled multiple rooms. Today an Apple Watch packs more power. We didn’t use fewer chips—we invented new uses for them. The same logic applies here. AI, autonomous vehicles, medical diagnostics, robotics—every leap demands more secure, always-on computing. “We’re definitely not in the bubble and I don’t think we’re going to see a curve going down in our lifetime and probably not our children’s lifetime,” he said.
Some smaller developers may fail and get gobbled up—the normal course of events in the emergence of new technology. But the next player will snap up the physical infrastructure.
China is experimenting with submerged coastal data centers for built-in cooling. Innovation keeps moving forward. The dot-com bubble was about stock prices and failing companies, not the underlying technology. The servers and fiber are here to stay. Every time we make computing more efficient, we find new ways to use it.
That’s not hype; that’s history repeating itself in real time.
None of this erases legitimate local concerns. Zoning still matters. Older plants need upgrades. Rural counties must avoid becoming Loudoun 2.0.
The Utah controversy illustrates a wider tension. Data centers are strategically vital for U.S. technological leadership, economic growth, and applications ranging from medical research to education and productivity. At the same time, communities have every right to demand rigorous environmental reviews, transparent water and power planning, and genuine local benefits.
O’Leary’s allegations of CCP-linked disinformation add a geopolitical dimension. We know the CCP has engaged for decades in influence operations in U.S. discourse, what they call fifth-generation warfare.
The blanket opposition peddled online ignores the measurable economic lift, the quieting of modern facilities, and the self-funding power strategy already underway. The genie is not going back in the bottle. The choice is whether Virginia shapes the future with facts or lets recycled myths from Loudoun County dictate policy.
Data centers driving up energy costs and fear of job losses have become major public concerns.
The American rebellion against AI is gaining steam
Amrith Ramkumar, Katherine Blunt, Lindsay Ellis
…
Delivering a commencement address at the University of Arizona, Schmidt told students the “technological transformation” wrought by artificial intelligence will be “larger, faster and more consequential than what came before.” Like some other graduation speakers mentioning AI, Schmidt was met with a chorus of boos.
…
Consumers resent energy-price jumps exacerbated by the spread of data centers. Workers fear widespread job losses. Parents worry about AI undermining education and harming children’s mental health. In recent months, the wave of anger has brought protests, swayed election results and spurred isolated acts of violence.
In April, a 20-year-old Texas man allegedly threw a Molotov cocktail at OpenAI Chief Executive Sam Altman’s home and made threats at the company’s San Francisco headquarters, according to a federal complaint filed against him. A few days earlier, someone fired 13 shots at the front door of an Indianapolis councilman who had recently approved a data center.
…
After bubbling up in a handful of races last year, it has exploded onto the ballot across the country. Voters in Festus, Mo., ousted four city council members a week after they approved a $6 billion data center. Dozens of communities in states from Maine to Arizona are trying to ban new data centers. Some 360,000 Americans are in Facebook groups opposed to the facilities, roughly quadruple the number from December, figures from organizations fighting the AI build-out show.
The point is these are legitimate concerns – AI is and will continue to cause job losses. Data centers are driving up power prices. But only one of these problems is fixable.
The data center electricity price rises are a symptom of a broken energy market, not a problem caused by data centers. Any rise in energy intensive economic activity, including a rise in factory jobs, would have caused similar upward pressure on energy prices. The problem is the market is not responding to the price signal. Higher power prices should be a signal to increase supply, but something is preventing market participants from responding adequately to changes in demand.
The source of market friction is likely red tape. For example, Data Centers could be built with co-located modular nuclear power plants, so they don’t ruin anyone’s power bill. Such an arrangement would bring down prices and improve local grid reliability, by allowing the data center to support the local grid with their in-house nuclear reactor during data center quiet periods. But imagine trying to push that solution through today’s planning approval processes.
The problem which can’t be fixed is job losses. Job losses are sadly inevitable – restricting or banning data centers in the USA won’t stop this, because China and Asia will keep pushing forward regardless of what the USA does. China with its surplus of electricity generation capacity already has a significant advantage in the AI race.
Much of the USA’s economic and military advantage is because of the USA’s enormous global trade network. Trying to stop data centers in the USA because of their impact on jobs will destroy the USA’s global competitive advantage. If the USA turns its back on data centers and AI, economic rivals which embrace AI will be in a position to destroy the USA’s competitive advantage and usurp the trade networks which propelled the USA to become the world’s dominant superpower.
Retraining might help mitigate the impact of job losses – and I’m not talking about trite non-solutions like President Biden’s advice to “learn to code”. There are plenty of jobs AI can’t do, and won’t be able to do well for the foreseeable future. Any job which involves complex and unpredictable environments, such as trade jobs like gas fitters, plumbers and electricians, will resist AI encroachment for a very long time. There will be more robotic tools for such jobs, but such jobs will still need humans in the loop to analyse the problems and figure out the solutions, and make sure the robots don’t do anything stupid.
Of course if you can “learn to code”, the jobs of people who serve the AI – data scientists, top tier programmers and expert maintenance engineers – will have real staying power.
My point is the problem is real, but banning or restricting data centers is not a viable solution. We can’t turn the clock back on today’s technological advances, any more than textile workers in the 1800s could stop machine driven mass production of woven cloth. All that can be done, and should be done, is to act early to cushion the blow, to help people distressed about job insecurity to find new jobs which the AIs can’t steal, and to figure out how to fix the regulatory roadblocks which are causing electricity price misery for anyone who lives next door to new data centers.
Update: A few people have questioned whether coding jobs have staying power, there is a lot of media hype about coding jobs disappearing in the near future.
In the AI world, coding is a lot more resilient than people assume. Coding is the art of explaining to computers how to do stuff it cannot figure out for itself. If the AI enhanced machines no longer need coders and can figure out everything, nobody will have a job, and we are not there yet. But how we code is changing significantly.
“Unprecedented surge in demand driven by the rapid expansion of large-load data centers and broader economy-wide electrification,” is noted by PJM as one of the the three drivers of grid strain, alongside “the accelerated retirement of dispatchable generation due to environmental policy and economics; and significant supply chain and permitting frictions that have extended the time required to bring new resources online.”
Data centers are the primary drivers of rising electricity consumption, according to the U.S. Energy Information Administration (EIA). The EIA expects electricity demand to grow by 1% in 2026 and 3% in 2023, potentially outpacing new energy supply.
“The result is a transition from an era of managing surplus to an era of managing scarcity,” the memo continues, highlighting long lead times on new energy developments. “Construction timelines have doubled,” the report continues. “Today, a new natural gas turbine plant – the reference technology for the capacity market – requires at least four years under optimistic assumptions from financial investment decision to commercial operation.”
Permitting requirements contribute considerably to long development times. The average environmental review takes over fouryears to complete, according to the Council on Environmental Quality.
PJM’s white paper highlights three paths toward keeping the grid solvent.
Path A aims to stabilize markets “through long-term forward commitments either through mandatory Load Serving Entity (LSE) hedging requirements or through a PJM-administered, long-term procurement, such as a tiered, multiyear capacity market.”
LSE is a catchall term for any authorized seller of electricity to its principal users.
Alternatively, PJM’s Path B institutes differential reliability.
“This could be implemented geographically (different states or zones procuring different levels of reliability) or by customer class (for example, with residential and native loads insulated from curtailment while unbacked new large load additions are curtailed first),” the memo notes.
Under the differential reliability model, data centers using preexisting supply would be subject to throttling during periods of high demand. Residential consumers who contributed to existing supply would be prioritized for output.
Path C aims to partially transition grid use to a supply-and-demand pricing model, while maintaining long-term contract optionality. PJM would “pursue a deliberate, phased shift of revenue recovery from the capacity market” combined with “long-term forward energy contracting requirements to protect consumers from increased energy price volatility and to support investment.”
“Constrained supply” is at the heart of PJM’s complaints. This shortage largely stems from the many regulatory reviews required to break ground on energy developments, according to the Consumer Energy Alliance.
The Trump administration has employed statutory actions to cut regulation and expedite permitting, notably overturning the Obama-era endangerment finding that allowed for federal regulation of carbon emissions.
These actions have not been made law by Congress. Though the House of Representatives passed the Standardizing Permitting and Expediting Economic Development Act in December 2025 to modify the National Environmental Policy Act, the legislation is frozen in the Senate.
All content created by the Daily Caller News Foundation, an independent and nonpartisan newswire service, is available without charge to any legitimate news publisher that can provide a large audience. All republished articles must include our logo, our reporter’s byline and their DCNF affiliation. For any questions about our guidelines or partnering with us, please contact licensing@dailycallernewsfoundation.org.
The “climate guilt” narrative, which has long emphasized reducing energy use, curbing consumption, and prioritizing rapid decarbonization through intermittent renewables, is slamming into the physics of the AI era: massive, always-on computational demand that doesn’t tolerate blackouts, brownouts, or weather-dependent supply.
The “climate guilt” narrative, which has long emphasized reducing energy use, curbing consumption, and prioritizing rapid decarbonization through intermittent renewables, is slamming into the physics of the AI era: massive, always-on computational demand that doesn’t tolerate blackouts, brownouts, or weather-dependent supply.
Global data center electricity use sat at ~415 TWh in 2024 (1.5% of world totals).
Projections have held firm or ticked higher: IEA’s Base Case still sees it doubling to ~945 TWh by 2030, with AI-accelerated servers (GPUs for training + inference) driving ~30% annual growth and nearly half the net increase.
Gartner pegs 448 TWh for 2025 rising to 980 TWh by 2030, while Goldman Sachs recently revised upward to 220% growth vs. 2023 levels—pushing global DC demand toward 1,350 TWh.
In the US (home to 45% of global DC load), consumption was 176 TWh in 2023 (4.4% of national electricity).
Forecasts now range 325–580 TWh by 2028 (6.7–12%) and potentially 606 TWh by 2030 (~12%), with data centers accounting for nearly half of total electricity demand growth through 2030.
Hotspots like Virginia could hit 41–59% of local supply. Utilities are warning of 49 GW shortfalls by 2028 in some models; hyperscalers are already delaying projects or self-generating with gas backups.
Inference (not just flashy training runs) now dominates long-term load—think everyday AI queries, agents, and enterprise deployments.
Efficiency gains in chips and cooling are real, but they haven’t offset the volume explosion.
Result: US power demand growth has flipped from flat to 1.7–2.5% annually, with gas filling near-term gaps per EIA.
The old framing—treat any new demand as sinful, push intermittents as the sole path, shame “wasteful” compute—hits hard limits when you need always-on, dense power at hyperscale.
Renewables have scaled impressively and will cover chunks of incremental load via PPAs, but:
Physics of reliability: AI clusters can’t flicker with the wind or sun. Overbuild + storage helps on paper, but transmission queues, land use, and backup needs (often gas) blunt the “pure green” story.
Growth mismatch: Total electricity demand is reaccelerating after decades of stagnation. AI is the biggest new driver, alongside EVs and industry. Sole reliance on variable sources risks blackouts, higher costs, or delayed builds—exactly what we’re seeing in 2025–26 grid congestion reports.
Big Tech’s own net-zero pledges are under pressure; some are quietly extending timelines or adding flexible “bridges.”
Emissions intensity per query keeps dropping, but absolute demand is the story.
This isn’t anti-climate—it’s recognizing that abundance (cheap, reliable watts) historically drives the innovation that cleans things up faster.
Hyperscalers aren’t waiting for narrative alignment—they’re voting with balance sheets and power purchase agreements.
As of April 2026:
Microsoft + Constellation: 20-year deal to restart Three Mile Island Unit 1 (now Crane Clean Energy Center, 835 MW). $1.6B project got a $1B Trump DOE loan in Nov 2025. Target online late 2027/2028, though transmission delays are the current bottleneck. Direct behind-the-meter power for AI data centers.
Meta: Massive Jan 2026 push—6.6 GW of nuclear projects via deals with TerraPower (up to 690 MW), Oklo (1.2 GW campus in Ohio), and others. Also extending Clinton plant life and eyeing restarts. Meta joined the “triple nuclear by 2050” pledge.
Amazon: Scaling SMRs hard—agreements with X-energy for multiple modules (aiming 5+ GW by 2039), plus Talen Energy PPA (1.9 GW existing nuclear) and Dominion. Behind-the-meter focus for campuses.
Google: Kairos Power SMR fleet—first online target 2030, more by 2035. World’s first corporate multi-SMR purchase agreement.
All four signed the World Nuclear Association pledge in 2025 to triple global capacity.
SMRs appeal for speed and modularity; restarts leverage existing sites.
Gas remains the flexible short-term fill, but nuclear is the firm, near-zero-carbon backbone AI craves. Geothermal and advanced options are in the mix too.
This isn’t greenwashing—it’s engineering reality.
AI itself will help: optimizing grids, accelerating fusion R&D, discovering better materials.
The economic upside (productivity, competitiveness) is too massive to subordinate to scarcity ideology.
Policymakers are noticing—permitting reforms, grid upgrades, and all-of-the-above supply are back on the table.
The tipping point favors outcomes over slogans: measure reliability, cost, emissions, and innovation, not guilt.
We’re shifting from penance to power abundance, and that’s how civilization (and cleaner tech) actually advances.
As of mid-April 2026, nuclear momentum for powering the AI age has shifted from hype to concrete deals, funding, and policy tailwinds—driven squarely by hyperscalers’ need for firm, 24/7, high-density power.
Big Tech isn’t just talking; it’s underwriting restarts, funding SMR fleets, and signing multi-gigawatt PPAs.
This directly counters the intermittency limits of renewables in the face of surging data center demand.
Small Modular Reactors (SMRs) are factory-fabricated nuclear power plants, typically under 300 MW per module (often 50-300 MW), designed for mass production, easier transport, and flexible deployment.
They contrast with traditional gigawatt-scale reactors by offering shorter construction times, lower upfront capital per unit, smaller land footprints, and enhanced safety via passive systems.
Many designs use light-water, gas-cooled, molten salt, or fast reactor technologies.
Black & Veatch executive Phil Fischer wrote this piece late last year making a case for natural gas for power generation, whether for the entire data center or for backup. In a free market, natural gas combined cycle, cogeneration, and peaking turbines have proven to be technologies of choice. With wind, solar, and battery subsidies on the chopping block, a new era for gas-fired electrical generation could be ahead.
Black & Veatch executive Phil Fischer wrote this piece late last year making a case for natural gas for power generation, whether for the entire data center or for backup. In a free market, natural gas combined cycle, cogeneration, and peaking turbines have proven to be technologies of choice. With wind, solar, and battery subsidies on the chopping block, a new era for gas-fired electrical generation could be ahead.
“Companies developing data centers are in a new era in which the scope of project planning must expand far beyond the data center itself and into the supporting physical infrastructure – power generation, water delivery and water treatment,” Phil Fisher begins. “Delays in this physical infrastructure have led to new ideas that open the window for options that previously had not been on the table.”
He provides “five important rationales” for natural gas (reproduced verbatim):
Significant cost savings for natural gas. Data center operators can save 38 to 45 percent of the cost to generate the same amount of electricity when considering the fuel source of natural gas versus diesel fuel. Although natural gas has a higher burn rate than diesel, the cost of natural gas is much lower, allowing for the substantial savings. Such cost efficiencies would be especially beneficial to hyperscale facilities requiring many large generators. The expectation of continued low pricing for natural gas is a major advantage going forward.
Multiple utilization and monetization options with natural gas. Diesel generators typically are used to provide backup power to data centers and only deployed in a power emergency or outage. Natural gas-powered generators can be used in a variety of ways beyond just as backup power. Depending on market location and utility regulations, owners could sell excess power back to the utility or to the spinning reserve wholesale power market. They’d also have the option to use natural gas to peak shave, thereby reducing demand charges. Natural gas also could be used to provide grid emergency response services. These monetization options are available whether the natural gas generation is for supporting regular operations or for backup.
Substantially cleaner fuel emissions and more versatility. Natural gas burns cleaner with lower emissions of carbon dioxide and nitrous oxide. Conversely, diesel emits more sulphur, nitrogen oxide and carbon dioxide. Natural gas also offers a higher potential for successful air permit approvals, and it provides for less noise pollution than diesel generators. It is stored in pipelines and is readily available on demand. Generally speaking, diesel fuel is only usable in storage for up to 24 months, meaning an operational maintenance plan is required to optimize performance.
Rapid progress in natural gas efficiency. There’s little doubt natural gas will become an even cleaner and more efficient resource in the future. Today, combined-cycle natural gas plants can achieve efficiency rates of 50 to 60 percent – and expected to exceed 65 percent within a few years. Advanced turbines operate with 30 to 50 percent hydrogen fuel content, and conversion to more efficient models is expected to advance, making them more cost-efficient. Operating with a higher hydrogen content will achieve cleaner operations. Eventually, that hydrogen content could reach 100 percent, with the ultimate goal of emitting nothing but water in order to comply with 2050 net-zero goals.
New focus on futureproofing. The idea of futureproofing is anticipating ways to keep assets useful, relevant and adaptive for productive use far into the future, helping ensure that today’s investments don’t become tomorrow’s stranded assets. One technology that could help safeguard natural gas relevancy decades from now is carbon capture, utilization and sequestration (CCUS). CCUS is projected to have the ability to capture and harness 90 percent of carbon dioxide emissions and produce carbon-based products of value. Certain industries are showing significant interest in this development today, and they are anticipated to embrace and put the technologies to work by 2030. This will signal a new direction for utility companies. Economy of scales could be achieved by 2040 that will allow for large-scale CCUS retrofits at natural gas plants. This exciting technology is actively in the experimental or pilot stages, and its development and progress will only advance.
He concludes:
Compared to other fossil fuels, natural gas offers superior efficiency rates, lower emissions and an abundant, reliable supply. Its lower carbon dioxide per unit of electricity is a huge plus for reaching green, sustainable goals.
As data center managers grapple with utilities being challenged to meet their power needs for new projects, they are having to think outside the box and look at self-generation of power or constructing their own substations. Whether used for power generation or backup, natural gas can be the perfect technology to ensure grid reliability.
——————————–
Phil Fischer handles new business for Black & Veatch in power generation. He has more than 25 years of experience in the distributed generation and data center industry working at a number of companies. Black & Veatch is a 100-percent employee-owned global engineering, procurement, consulting and construction company was founded in 1915.
You must be logged in to post a comment.