Friday, November 28, 2025

 

Ontario Clears OPG to Launch $26.8B Pickering Nuclear Refurbishment

Ontario has given Ontario Power Generation the formal go-ahead to begin executing the refurbishment of the Pickering Nuclear Generating Station’s Units 5–8, marking one of Canada’s largest energy infrastructure commitments in decades. The decision clears OPG to advance detailed planning ahead of project execution in early 2027.

The province’s approval enables OPG to start transitioning Pickering into its next operational phase, securing 2.1 gigawatts of low-carbon baseload power for more than 30 additional years. That represents enough electricity to supply over two million homes and will help Ontario meet steeply rising electricity demand tied to industrial growth, electrification and population expansion.

The green light allows OPG to begin the definition phase of Pickering’s refurbishment through 2026, leveraging over 7,000 lessons learned from the Darlington Refurbishment and the utility’s ongoing small modular reactor (SMR) program. Construction and component replacement work will begin in 2027, with completion expected in the mid-2030s.

With an estimated all-in cost of $26.8 billion, the Pickering overhaul ranks among the largest infrastructure projects in the country. According to the Conference Board of Canada modelling, the project is projected to add $38.2 billion to Ontario’s GDP and $41.6 billion nationally over its lifespan. Roughly 85–90% of spending will remain in Ontario, bolstering the province’s nuclear supply chain and creating substantial regional economic benefits.

Employment impacts are similarly significant: refurbishment and continued operations are expected to support an average of 7,500 jobs annually across Canada, including 30,500 jobs per year during the refurbishment phase.

Major contracting has already begun, including a $2.1-billion award to CanAtom for early engineering and procurement tied to the retube, feeder and boiler replacement program.

Pickering’s four-unit overhaul mirrors elements of Darlington’s multi-year life-extension—such as replacing 380 fuel channels per reactor—but adds considerable complexity. Unlike Darlington, Pickering’s project includes the replacement of all 48 steam generators, a first-of-its-kind scope for Ontario’s CANDU fleet.

The project also includes construction of a 1.5-kilometre deep-water intake system to secure colder cooling water—an infrastructure element Pickering lacks from its original design.

Pickering holds an operating licence valid through 2028. OPG has applied for a new 10-year licence from the Canadian Nuclear Safety Commission that would cover refurbishment activities and post-refurbishment operations.

Ontario is in the midst of the largest nuclear build-out in Canadian history. Alongside Darlington’s completed and ongoing refurbishment stages, the province is advancing the G7’s first grid-connected SMR at Darlington, expanding the Bruce Nuclear site, and now embarking on Pickering’s full life-extension. Together, these moves are designed to ensure long-term grid reliability while meeting aggressive climate and electrification targets.

By Charles Kennedy for Oilprice.com


Pickering refurbishment gets government go-ahead



The Ontario government has approved Ontario Power Generation's plan to refurbish four Candu nuclear reactors at Pickering, clearing the way for a start to the execution phase of the project, subject to final regulatory approvals.

Inside the turbine hall at Pickering (Image: OPG)

"For more than 50 years, nuclear power has been the backbone of Ontario's energy grid, with the Pickering Nuclear Generating Station as one of the best-performing and continuously operating nuclear stations in the world," Minister of Energy and Mines Stephen Lecce said as he announced the government's approval.

"To deliver on our major jobs plan to get Canadians working, extending the life of the facility will create jobs for tens of thousands of skilled workers while we build up the Ontario supply chain and keep 90 per cent of our project spend in this country."

Pickering units 5-8 - known as Pickering B - began operations in the mid-1980s and had been scheduled to end electricity production this year, but in 2022 the provincial government directed OPG to keep them in operation until 2026 and to reassess the feasibility of refurbishing the units. The government gave OPG the go-ahead to begin the initiation phase of the refurbishment project in January 2024, and one year later, it gave its permission for the start of the project definition phase.

The final budget for the Pickering refurbishment is CAD26.8 billion (USD19.1 billion), but according to a study by the Conference Board of Canada, the mega-project - which will enable the plant to continue in operation for up to 38 more years - will increase Ontario's GDP by CAD38.2 billion in 2024 dollars over the project's lifespan, including CAD17 billion during the refurbishment phase, and increase Canada's national GDP by CAD41.6 billion over its lifespan.

Some 30,500 jobs will be created during refurbishment, while the station's operation will sustain 6,700 jobs. To reduce costs, the government says it will continue exploring potential financial instruments that would benefit ratepayers, while OPG "continues to explore other optimal financing arrangements".


Pickering (Image: OPG)

Refurbishment of CANDU reactors is a major undertaking in which thousands of parts - including major components such as steam generators and pressure tubes - are inspected or replaced before the reassembled reactor returns to service. OPG has been carrying out refurbishment of four units at its Darlington site, and thousands of lessons learned there will be incorporated at the Pickering refurbishment which the company said is a more complex and larger project than Darlington. The work at Pickering will include replacing all 48 steam generators across the four units - components that were inspected rather than replaced at Darlington.

With the green light from the government, OPG said it will continue to advance the definition phase, which will last through much of 2026. It has already signed contracts for the refurbishment with Ontario-based suppliers, including a CAD2.1-billion contract with CanAtom for early engineering and procurement to prepare for the execution of the project's Retube Feeder and Boiler Replacement Program. It expects to begin the Project Execution Phase in early 2027, with completion expected by the mid-2030s.

"This is a truly remarkable moment for Pickering Nuclear, our employees, and the community," OPG President and CEO Nicolle Butcher said. "We are ready to leverage all of our ongoing lessons and experience from our Darlington Refurbishment and the construction of the G7's first Small Modular Reactor to deliver another successful nuclear refurbishment for Ontario. Through this project, we will ensure this important station continues to play a key role in our future - by generating critical energy and economic benefits, sustaining high-quality jobs, and ensuring Ontario and Canada's energy security."

Pickering units 1 and 4 - the last of the four units making up Pickering A - ceased commercial operations in 2024.

Britain’s Nuclear Reform Set to Stall Over Legal Concerns


  • Labour is reportedly pausing nuclear-sector reforms despite a sweeping report urging planning and regulatory changes to cut costs and accelerate new projects.

  • Legal concerns raised by a government adviser have prompted Reeves to withhold the recommendations from the upcoming Budget, delaying growth-focused measures.

  • Industry leaders, MPs across parties, and think tanks argue that fully adopting the review is essential to lowering energy costs, boosting investment, and reviving UK economic growth.

The Labour government is set to hold fire on pushing through sweeping reforms to nuclear energy due to a legal adviser’s concerns over the “UK’s environmental, trade and human rights obligations”, it has been reported. 

The Innovate UK member John Fingleton published a damning report on red tape blocking the expansion of the country’s nuclear industry on Monday.

His extensive report prompted calls from across the political spectrum for Rachel Reeves to accept recommendations easing planning rules and lowering costs for investors. 

But ITV News has now reported that the Chancellor will not include the growth-focused recommendations in her Budget speech on Wednesday. 

The broadcaster reported that the Chancellor will make reforms “subject to further work and review” after a government adviser voiced concerns about the legal crossovers in the paper with UK obligations. 

The landmark report on nuclear energy underscored the need for radical changes to the planning and regulatory system to allow more plants to be built to support the UK’s drive to lead on AI. 

Fingleton outlined how £700m was being spent by Hinkley Point C bosses on protecting one salmon every ten years in its “fish disco” project that uses noise to keep animals away from the plant in Somerset. 

His report said investors and builders had to deal with a “fragmented” system of environmental or community regulators and red tape to get projects over the line, leading the UK to be the “most expensive” place to build nuclear power. 

Labour government urged to back growth

MPs from different parties, growth activists and think tanks have joined up to call for 28 recommendations in the report to be accepted by the Chancellor at the Budget. 

A letter signed by Labour Growth Group chair Chris Curtis and groups including the Centre for British Progress, Tony Blair Institute and the YIMBY Alliance said: “Britain has the chance to once again lead the world in building nuclear power, which is the backbone of industrial communities and jobs across the country, and also at the heart of our climate ambitions.

“There is a once in a generation opportunity to build the energy and industrial infrastructure of the future on these islands.”

Reform UK’s Zia Yusuf and the Conservative Party’s Kemi Badenoch also spoke out in support of the review. 

Lawrence Newport, the head of campaign group Looking for Growth, told City AM: “Such a significant shift would set Britain on a clear path towards affordable energy and, consequently, economic growth.

Indeed, reducing energy prices is a prerequisite for any government hoping to reverse decades of stagnation and decline. With cheaper energy, British businesses could start and scale here uninhibited.

Rachel Reeves must make a decision. If the Chancellor really is as serious about growth as she says, if she actually wants to make our energy cheap again, she will accept all of the Nuclear Regulatory Review’s recommendations.”

By Mauricio Alencar via CityAM


 World Nuclear News


GDES, KEPCO KPS team up for Spanish and French markets

Spanish energy services group Grupo Dominguis Energy Services and South Korea's KEPCO Plant Service & Engineering have signed a collaboration agreement to further their joint development and expand cooperation in the Spanish and French markets, specifically in the field of nuclear power plant maintenance.
 
The signing of the agreement (Image: GDES)

The agreement was signed on 17 November at Korea Electric Power Corporation's headquarters in Naju-si, South Korea, in a ceremony led by Ho-Kwang Jun, Executive Vice President of KEPCO Plant Service & Engineering (KEPCO KPS), and Héctor Dominguis, CEO of Grupo Dominguis Energy Services (GDES). 

GDES said the alliance will enable both companies to share know-how, resources and capabilities "in pursuit of common objectives". It said the partnership "brings together GDES's long-standing expertise in the European nuclear sector and KEPCO KPS's strong technological capabilities as a leading South Korean provider of operation, diagnostics and maintenance services for energy facilities".

KEPCO KPS is a South Korean public company, operating in 25 countries and employing around 5,800 people. The company is widely recognised for its technical and operational excellence and specialises in diagnostics, performance improvement and operation and maintenance (O&M) services for both conventional and renewable energy facilities. Its technology-driven approach is supported by dedicated R&D centres focused on digitalisation, predictive maintenance and advanced solutions for the management of critical assets.

GDES has more than 45 years of experience in the nuclear sector. The group has participated in the construction of Spain's nuclear power plants, as well as in their maintenance and decommissioning, "becoming a European benchmark with a presence in projects in Spain, France, Sweden, Switzerland (CERN), the United Kingdom and Italy".

"Through this alliance, GDES strengthens and broadens its innovation, technology and O&M capabilities, reinforcing its value proposition for the Spanish nuclear sector while accelerating its strategic expansion in international markets," the company said.

Dutch support for molten salt reactor demonstrator facility

Molten salt reactor developer Thorizon announced it has secured broad industrial and regional support in the Netherlands for the construction of Thorizon Pioneer, a nuclear demonstrator for next-generation molten-salt reactors.
 
(Image: Thorizon)

The announcement was made at the Made for Nuclear event, where Thorizon and a wide network of partners signed a joint Ambition Agreement to accelerate the project. The coalition brings together leading Dutch industrial companies including VDL Groep, nuclear operator EPZ and engineering and construction firm Haskoning, supported by a strong ecosystem of universities, knowledge institutes, manufacturing, energy and high-tech service providers. Several Dutch provinces have confirmed their engagement and are now assessing potential sites for the demonstrator, as well as opportunities to establish an innovation hub around the facility with testing capabilities for materials and components.

The Thorizon Pioneer is part of Thorizon's roadmap to industrialise molten salt reactors (MSRs). The facility will be a smaller, non-commercial version of the Thorizon reactor focused on validating the reactor core without electricity production. The Thorizon Pioneer will be the first nuclear molten salt installation operating in Europe, the company said, providing unique opportunities for research and development. 

"It is designed to demonstrate Thorizon's ability to design, license and build a nuclear installation," Thorizon said. "The Pioneer will also enable the qualification of supply chain partners, testing of key materials and components, which is central to lowering construction and operating costs in the full-scale reactor."

"The Netherlands has a great and entrepreneurial manufacturing and services industry, skills we intend to combine with the R&D expertise and manufacturing expertise in nuclear," said Thorizon CEO Kiki Lauwers. "We can build on the knowledge of the Pallas medical isotope reactor and ongoing research at TU Delft and Differ. We hope to continue the strong cooperation in Europe to make this project a success and thank our partners and the provinces involved for this opportunity."

MSRs use molten fluoride salts as primary coolant, at low pressure. They may operate with epithermal or fast neutron spectrums, and with a variety of fuels. Much of the interest today in reviving the MSR concept relates to using thorium (to breed fissile uranium-233), where an initial source of fissile material such as plutonium-239 needs to be provided.

To prepare for European licensing, Thorizon said it is engaged in a joint preparatory safety review with the Dutch, French and Belgium nuclear safety authorities. The next major milestone is securing a suitable site. Over the coming months, Thorizon will work with several provinces and others to evaluate potential locations, ideally one already licensed for research-reactor activities. In parallel, the company is advancing the design of the Pioneer, expanding engineering collaboration with industrial partners and finalising the financing strategy for the demonstrator.

The project will be supported through a combination of private investment and public funding. Thorizon is backed by Positron Ventures, Invest-NL and the regional development agencies of Zeeland and North Holland, and has secured grants from the French government, the European Commission's Joint Transition Fund in Zeeland and the Province of Brabant. The initiative is also prioritised by the European Commission within the EU SMR Alliance.

Thorizon - a spin-off from NRG, which operates the High Flux Reactor in Petten in the Netherlands - is developing a 250 MWt/100 MWe molten salt reactor, targeted at large industrial customers and utilities. The molten salt fuel adopted by Thorizon uses a combination of long-lived elements from reprocessed used nuclear fuel and thorium. The reactor will be able to recycle long-lived waste from existing nuclear facilities. The Thorizon One concept is unique in that the core is composed of a set of cartridges that is replaced every five to ten years. This, the company says, overcomes two molten salt design obstacles: material corrosion and handling of used fuel volumes.

The company says it is conducting pre-feasibility studies at three nuclear-designated sites in France, Belgium and the Netherlands, targeting construction by 2030.

Equinix signs up for power from first Stellaria reactor


French molten salt reactor developer Stellaria has signed a pre-order agreement with data centre developer and operator Equinix. Under the agreement, Equinix has secured the first power capacity reservation on the Stellarium, the reactor that Stellaria plans to deploy starting in 2035.
Stellarium, the first Breed and Burn reactor covered by the pre-order agreement (Image: Stellaria)

The agreement between Stellaria and Equinix is part of several Equinix initiatives in the field of alternative energy. In August, the company announced collaborations with five energy providers, including Stellaria in France, to support the growth of its AI-ready data centres.

The Stellarium reactor proposed by Stellaria - a spin-off from the French Alternative Energies and Atomic Energy Commission (CEA) - will be very compact (measuring 4 cubic metres) and will be able to use a diversified range of nuclear fuels (uranium, plutonium, MOX, minor actinides, even thorium). Stellaria says the reactor is "the world's first reactor to operate with a liquid fuel capable of destroying more waste than it produces".

"We chose Stellaria because it is one of the few companies in the world capable of making our high-performance AI data centres energy resilient, while combining high security and flexibility," said Régis Castagne, Managing Director, Equinix France.

"Their Stellariums natively generate 250 MWe of clean energy and uniquely consume their own waste, eliminating the need for on-site storage and addressing a major operational challenge. This breakthrough will allow Equinix to operate with secure, carbon-free energy 24/7. It represents a major step forward in building the next generation of sustainable digital infrastructure."

California-headquartered Equinix operates more than 270 data centres in 77 major metropolitan areas around the world. Equinix data centres are currently covered by 100% renewable energy in Europe, and the company is investing significantly in renewable energy power purchase agreements in France. The company is now expanding its energy portfolio to support the development of reliable and sustainable nuclear electricity to power its high-performance AI data centres with Stellaria.

"The signing of this first contract is an important milestone for Stellaria. It sends a strong signal to the sector and proves that our roadmap is credible, with the deployment of our Stellariums planned for 2035. The digital sector is fully engaged in the sustainability to reduce environmental impact: this contract lays the foundation for lifetime energy autonomous data centres, and we are proud to support a major player like Equinix in this project," said Stellaria CEO Nicolas Breyton.

Molten salt reactors (MSRs) use molten fluoride salts as primary coolant, at low pressure. They may operate with epithermal or fast neutron spectrums, and with a variety of fuels. Much of the interest today in reviving the MSR concept relates to using thorium (to breed fissile uranium-233), where an initial source of fissile material such as plutonium-239 needs to be provided. There are a number of different MSR design concepts, and a number of interesting challenges in the commercialisation of many, especially with thorium.

Kyrgyzstan considering possible Russian SMR plant

Kyrgyzstan is exploring the possibility of building its first nuclear power plant using Russian small modular reactors, Vladimir Putin has said.
 

(Image: Kremlin.ru)

In a statement after talks with Kyrgyzstan's President Sadyr Japarov, the Russian President said: "Energy is one of the most important sectors of Russian-Kyrgyz cooperation. Our country fully meets Kyrgyzstan's gasoline and diesel needs, and does so on preferential terms, without levying export duties, which also has a direct economic impact on the republic."

He also said Gazprom supplies natural gas, Russian companies participate in the design and modernisation of hydroelectric power plants and joint construction of a large solar power plant is planned.

"Rosatom is implementing a large-scale reclamation programme for uranium mining sites. The possibility of constructing the republic's first nuclear power plant using advanced Russian small modular reactor technologies is being explored. These reactors, I emphasise, meet the most stringent safety and environmental protection requirements," he said.

In September, during the World Atomic Week international forum in Moscow, Altynbek Rysbekov, Deputy Minister of Energy of Kyrgyzstan, said the country was working with Rosatom "to determine which areas could [be] suitable for a nuclear power plant and what capacity it should have", according to the in-house Strana Rosatom publication.

In the published text of his speech after the talks, Kyrgyzstan's president did not refer directly to nuclear energy plans, but talked about the close links between the two countries and their future development: "Our government agencies have been instructed to accelerate joint efforts to supplement the bilateral agenda with new promising areas and identify new areas of mutual interest. Just now ... a new set of bilateral documents has been signed in the areas of education, healthcare, migration, security, economic and military-technical cooperation. I am pleased to add that the contractual and legal framework between the two countries includes more than 400 documents." 

Russia has developed land-based and floating small modular reactor (SMR) plants featuring the RITM-200N water-cooled reactor, which is adapted from nuclear-powered icebreaker technology, with thermal power of 190 MW or electric power of 55 MW and with an intended service life of 60 years.

The first export order has been signed with Uzbekistan for two of the units to feature alongside two large VVER-1200 units at the country's first nuclear power plant. The first SMR unit is scheduled to go critical in late 2029.

Electricity generation in Kyrgyzstan is mainly from hydro (over 90%). The country has no nuclear power, but in January 2022 signed a memorandum of cooperation with Rosatom for a feasibility study into the construction of an SMR plant based on the RITM-200N. The country has a history of uranium mining, but no uranium mining is carried out there today

 

The AI Gold Rush Is Redrawing America’s Energy Map

  • The U.S. data center pipeline has surged to 245 GW, dominated by mega-projects increasingly sited in Texas and designed around their own gas-fired power generation.

  • A handful of massive campuses distort capital flows and rely on unconventional financing methods, intensifying concerns about froth and an AI-driven infrastructure bubble.

  • These projects threaten higher long-term natural gas and electricity costs while posing new reliability challenges, likely prompting state intervention in power markets.

As industry concerns over an AI bubble mount, the scale of data center ambitions continues to grow. As of mid-October, the US data center pipeline reached 245 GW of planned capacity, driven by a handful of enormous, speculative projects. These projects, and the renewable deals hyperscalers are signing, skew heavily toward Texas. More than a quarter of pipeline capacity targets the state, whose pipeline nearly doubled from 35 GW in Q1 to 67 GW in Q3. 

Data Center Alley gives way to Data Center Prairie  

It is conventional wisdom that for data center developers, ‘access to power’ has become the mantra. That is, the importance of fiber proximity to end customers and other data centers has been superseded by the imperative to secure power from utilities. The giga-scale campuses that have been announced this year, sited in Pennsylvania, Wyoming, and particularly West and North Texas, signal yet another pivot in strategy. Developers have increasingly given up confidence that utilities can meet their power and timeline demand and instead seek to build their own generation based on local natural resources. 

Most frequently, this resource is natural gas, with a particular focus on the Permian in Texas. Pacifico Energy’s 5-GW GW Ranch, poolside’s 2-GW Project Horizon, and FO Permian’s 5-GW campus in Midland County are examples. The gas network in the US is at capacity, and building new gas generation far from supply means paying for and - perhaps more importantly for developers - waiting for new pipeline capacity. 

In some cases, it should be noted, campuses are being leveraged for their scale and solar or wind resource, such as Tract’s data center parks in Nevada and Utah, and Quantica’s Big Sky Digital Infrastructure campus in Montana. While more headlines have made about the potential for batteries to make data centers flexible assets, we’ve found that batteries are being planned much more commonly to balance renewables at large campuses and to add needed fast-ramping capacity to onsite gas. 

Distorting the capital landscape 

These mega campuses not only reflect a new siting strategy: they distort the capital landscape. The 2% of projects over US$17 billion represent 42% of overall capital deployment, with the 60% of projects below US$1 billion contributing only 8%. With questions already intensifying regarding froth in project investment and company valuations, these projects promise to drive even more froth. 

The two highest-cost projects, Project Jupiter in New Mexico (US$160 billion) and Project Kestrel in Missouri (US$100 billion), are an order of magnitude more expensive than the campuses being developed by companies like Meta and Microsoft, without appreciable increase in IT infrastructure. They are notably being funded through novel financial engineering: Their developers will be both the payers and payees of industrial revenue bonds (IRBs) issued by the local government, a contrivance that enables tax benefits. 

Betting big on onsite generation 

The new energy-chasing campuses are committing to onsite generation in a way that developers have been hesitant to in the past. Hyperscalers have been very clear that they prefer grid power, which requires less operating risk, shorter contract commitments, and no exposure to scope 1 emissions. Developer commitment to onsite generation reflects a bet either that hyperscalers urgency to deploy will overcome their concerns, or that the developer will be able to sell to a new generation of hyperscalers with greater risk tolerance and less concern for sustainability. 

The net result is that while the number of projects in the pipeline with onsite generation has increased only to 10%, these projects represent a whopping 34% of pipeline capacity.

Unsurprisingly, the vast majority of sites are in Texas, and the technology is gas turbines. 

This trend could have significant implications for both energy affordability and reliability. Whether on- or off-grid, projects with utility-scale gas generation will increase gas burns, competing with LNG exports and raising the long-term price of natural gas. This will drive up both gas and electricity bills across the country. To the extent to which turbine-based projects are off-grid, utilities will have an even harder time obtaining turbines given production limitations. That could pose reliability challenges for on-grid load growth, including electrification. 

Whether or not a bubble bursts, these affordability and reliability challenges are likely to provoke state intervention to protect customers. As that begins to happen, all bets are off. 

By Ben Hertz-Shargel via Zerohedge

 

The Cloud’s Not Fluffy, It’s a Hot, Loud Factory

  • AI’s shift from CPUs to ultra–power-dense GPUs has created a structural surge in electricity demand that is outpacing efficiency gains and overwhelming the grid.

  • The need for 24/7 baseload power is delaying coal retirements, boosting natural gas, and shifting data center growth to regions with cheap land, weak regulation, and vulnerable communities.

  • Global transformer shortages, mineral constraints, and interconnection bottlenecks mean that compute—not oil—may become the next strategic resource nations hoard.


The Cloud" might be the greatest branding trick in history. It sounds fluffy, ethereal, and notably light.

It implies that our digital lives…our emails, our crypto wallets, our endless scrolling…exist in some vaporous layer of the atmosphere, detached from earthly constraints.

But if you actually drive out to Loudoun County, Virginia, or stare at the arid plains of Altoona, Iowa, you realize the Cloud is actually just a very big, very loud, and very hot factory.

We’ve been telling ourselves a lovely story about the energy transition. We were retiring coal plants, building wind farms, and decoupling economic growth from carbon emissions. It was all going according to plan.

For years, the tech sector achieved relative decoupling…

Moore's Law kept server efficiency gains ahead of the curve, allowing internet traffic to surge while power demand grew slowly.

The exponential curve of AI, however, has shattered this delicate balance. AI workloads are so compute-intensive that demand is now skyrocketing faster than efficiency gains can compensate. 

This is a re-coupling with physics..and the defining narrative of the next decade isn't about supply anymore.

Now it's about a structural shift in demand that almost nobody priced in: The thermodynamics of Artificial Intelligence.

According to the International Energy Agency (IEA), global electricity demand from data centers is projected to more than double by 2030. This is the same as the entire annual electricity use of a country like Japan.

The invisible hand is hitting a concrete wall.

The question is no longer if the grid can handle it, but what is making the demand curve look like a rocket launch. The answer isn't better software or smarter algorithms; it's the raw physics happening inside a rack that now demands the power of a city block.

The Thermodynamics of "Thinking"

To understand why the grid is struggling right now, you have to look at the silicon.

For a long time, we ran the internet on CPUs (Central Processing Units). These are the general managers of the chip world. Efficient, predictable.

But Generative AI doesn't want a manager. It wants a battalion of mathematicians. It runs on GPUs (Graphics Processing Units), specifically monsters like Nvidia’s H100.

Here’s what that actually means for power draw:

  • Traditional server rack: Draws about 5 to 10 kilowatts (kW).
  • Modern AI rack (H100s/Blackwell): Draws 50 to 100 kW.

We have effectively moved from powering a toaster to powering a neighborhood, all inside the same metal box. Air cooling…fans blowing over hot metal…doesn't work anymore. Air just isn't physically dense enough to move that much heat away.

We are now plumbing data centers like chemical refineries, running liquid coolant loops directly to the silicon die.

This is the new reality of Direct-to-Chip (DTC) cooling. It is already happening in cutting-edge AI centers because it is the only way to manage the extreme heat density of chips like the H100.

Liquid cooling saves energy compared to air conditioning. While the chip itself still draws 100 kW, the overall cooling system…the pumps and chillers…consumes far less power than running massive air handlers for the whole room. This makes it an efficiency measure born of necessity.

The next step is Immersion Cooling, where entire server racks are submerged in a non-conductive fluid. This is also being deployed now, often in pilot programs and specialized facilities.

This shift from fans to specialized plumbing and chemically inert fluids is the physical realization of the industrialization of thought.

Just like the industrialization of textiles or steel, it requires massive inputs of raw power and exotic, specialty materials. This industrial intensity demands something traditional renewable sources…intermittent solar and wind…struggle to provide: reliability.

When an AI training run costs tens of millions of dollars, a 1% flicker is an existential threat. 

The Dirty Secret of the "Green" AI Boom

Every major tech CEO is currently on a podcast tour talking about their "Net Zero" 2030 goals. And sure, they are buying a lot of paper credits.

But physics doesn't care about carbon offsets. The reality is that AI needs baseload power. It needs to run 24/7/365 with "five nines" (99.999%) of reliability.

You know what provides that?

According to IEA data, coal still accounts for about 30% of global data center power. And in the U.S., natural gas is doing the heavy lifting, covering over 40% of demand.

The irony is palpable. We spent billions trying to kill coal, only to have the most futuristic technology on earth, AI, throw it a lifeline.

In places like Virginia or Kansas, utilities are delaying the retirement of coal plants. They simply cannot risk the grid instability when a gigawatt-scale data center comes online.

The "future" is being powered by the "past."

The need for this reliable baseload power, combined with the sheer gigawatt-scale hunger of these new facilities, is now fundamentally reshaping the American power landscape. Capital always flows to the path of least resistance—and right now, that path runs right through communities that have never seen a single dollar of tech prosperity.

The New Geography of Power (and Inequality)

This energy hunger is redrawing the map. We are seeing a "K-shaped" geography of infrastructure.

In the U.S., "Data Center Alley" in Northern Virginia supposedly handles 70% of the world's internet traffic. But the grid there is tapped out. You can't get a new hookup for years.

So, the capital is fleeing to places with looser regulations and cheaper land: Texas, Ohio, Arizona.

But this brings us to the friction point. These facilities are neighbors. And they are often bad neighbors. They are loud, they consume massive amounts of water for cooling, and they raise local utility rates.

There is also a significant Environmental Justice component here. Industrial infrastructure is rarely sited in wealthy neighborhoods.

According to the NAACP's "Fumes Across the Fence-Line" report:

  • African Americans are 75% more likely than white Americans to live in "fence-line" communities (areas adjacent to industrial facilities).
  • A disproportionate number of fossil-fuel peaker plants, which fire up when data centers max out the grid, are located in low-income areas and communities of color.

This directly contributes to higher rates of asthma and respiratory issues.

While the "invisible prosperity" of AI stock gains flows to portfolios in San Francisco and New York, the "visible decay"...the pollution, the water usage, the hum of the cooling fans…is localized in communities that often see none of the upside.

Even if a community were willing to bear the cost, the industrial machine that once smoothly supplied the electrical grid is choked.

The problem is no longer just where to put the data center, but how to physically connect the massive, power-hungry factory to the existing grid infrastructure. This process is crippled by a global bottleneck of essential, non-digital hardware.

The Great Transformer Shortage

Let’s say you have the money, the land, and the permits. You still have a problem. You can’t get the gear.

The lead time for a high-voltage power transformer used to be 12 months. Today? It’s 3 to 5 years.

We are trying to rebuild the electrical grid at the exact moment everyone else is trying to electrify cars and heat pumps. The supply chain is fractured.

We are also running out of the raw stuff: Copper. Lithium. Neodymium for the magnets in the cooling fans.

We are dependent on China for the processing of nearly all these critical minerals. As I explained in this “Data Center Guide,” we are realizing that the digital economy is actually a material economy.

If China restricts graphite or gallium exports (which they have started doing), the Cloud stops growing.

The "Trust Me, Bro" Efficiency Pitch

The counter-argument from Silicon Valley is the "Handprint" theory. The pitch goes like this: Yes, training the AI uses a lot of energy, but the AI will make the rest of the world so efficient that it pays for itself.

The IEA models suggest that AI could optimize logistics, manage smart grids, and reduce building energy usage by 10-20%.

And honestly? It’s a compelling argument. If AI can figure out how to drive a truck platoon 5% more efficiently, that saves more carbon than the data center emits.

But this is a long-term bet against a short-term, guaranteed withdrawal of power.

The core efficiency problem is two-fold:

  • Training vs. Inference: Training a colossal model takes a massive, months-long burst of power. The resulting AI is then put to work performing inference…answering questions. While inference is far cheaper per interaction than training, its global volume is exponentially growing, turning tiny energy costs into a massive, persistent drain.
  • The Hardware Treadmill: A high-end CPU might last 5-7 years in a data center. The new AI GPUs are considered obsolete in as little as two years. This brutal, accelerated hardware cycle…the constant replacement of power-hungry H100s with even more power-hungry Blackwells…means that the embodied carbon and raw materials tied up in the silicon are never given a chance to pay back their energy debt over a reasonable lifespan.

We are spending the carbon now in hopes of efficiency later. While the industry is working on "smarter" silicon, efficient ASICs for inference, that transition won't arrive fast enough to save the grid from the current exponential surge.

What Comes Next?

We are moving from an era of Generation Constraints to Connection Constraints.

The most valuable asset in the world right now isn't the H100 chip; it’s a signed interconnection agreement with a utility company. The "queue" to get on the grid is the new velvet rope.

This is going to force a few things:

  • Off-Grid AI: Tech giants will stop waiting for the utility. They will build their own SMRs (Small Modular Nuclear Reactors) or massive solar farms with battery storage, effectively taking their ball and going home.
  • Sovereign Compute: Nations will realize that "compute" is a strategic resource like oil. You will see countries hoarding power to feed their own AI models rather than exporting it.
  • The Efficiency Wall: We will hit a point where the cost of power makes brute-force AI training uneconomical, forcing a shift to "smarter" chips (ASICs) and maybe, eventually, neuromorphic or photonic computing.

The invisible hand is dealing cards, but the laws of thermodynamics are calling the bluff. The virtual world requires real power, and for the first time in a long time, we are realizing that "unlimited data" was a temporary illusion.

By Michael Kern for Oilprice.com