A Youtube video did the rounds last week trying to explain what the United States has achieved in Iran.

• In summary, the Strait of Hormuz, which wasn't closed until after the US attacked, is now being reopened as a victory.
• The regime change that was never the stated goal is being claimed as a win.
• Iran now has more money and more leverage over the strait than it did before the old deal was scrapped for being terrible.
• On Saturday, 21 hours of face-to-face peace talks in Islamabad collapsed. The US demanded Iran surrender its nuclear programme. Iran demanded control of the strait, war reparations and a ceasefire across Lebanon.
• Neither side moved and within hours, the US announced a full naval blockade of Iranian ports starting Monday morning. Oil jumped back above $100 a barrel.
• Iran's chief negotiator posted a photo of gas prices near Washington with the caption: "Enjoy the current pump figures."

Now, they say overhead walks on two legs, meaning it's people that cost the most money in a business. However, people don't flip-flap their fees like petrol stations.

So do expect AI companies to move to dynamic pricing if such incredible shocks to energy and key resources markets, such as precious metals, become more volatile. Given everything we are seeing, I think they will.

Your employee carries the risk of inflation over the year in the hope of a pay rise at some point. That's a slow game and while some may say that's unfair, it's how it's always worked. Checkpoints spread out.

But the most aggressive tech companies in the world will not carry the energy risk. They have no intention of doing so.

That is a challenge to your business because you've probably not baked in unstable energy prices in your risk modelling - if at all - to the factor they are about to display. In fact, I would go as far to guess that many companies still see energy as a relative constant.

So where does this all head next?

Professor Robert Pape, who spent 20 years running Iran war simulations for the White House, told Steven Bartlett on the Diary of a CEO that there is a 75% chance the US escalates to stage three: boots on the ground.

Then Professor Steve Keen, who predicted the 2008 crash, warned on the same show that food prices could double because 46% of global urea supply transits the same Gulf corridor now under blockade.

I watched all of this unfold while sitting at the kitchen table working out how much it costs to drive my sons to school. The school is two miles away (yes, we will be cycling now the weather is nicer).

The car is a diesel. At current UK diesel prices of £1.91 per litre, a four-mile round trip costs somewhere around £1.50 to £1.80. Over 190 school days, that is between £285 and £340 a year. Just to move a car two miles down the road and back.

The same trip in an electric car costs about 25 to 35 pence. Roughly a fifth of the price. A diesel doing 50 miles per gallon costs around 17 pence per mile. An EV charged at home on an overnight tariff costs two to seven pence.

Sourcing technology to turn strategy into reality

Proudly sponsors the Executive Summary

Find out more

March 2026 was the best month on record for UK electric car registrations, with battery electric vehicles up 24% year on year. Used EV dealerships are reporting sales increases of 60%, with queues out the door on a Saturday morning.

In Ireland, the consequences arrived faster and harder. Diesel rose 40 pence per litre in a single month, the largest increase on record. On 7 April, convoys of tractors, trucks and cars blockaded the M50 in Dublin, O'Connell Street and major motorways across the country.

By the weekend, 600 of Ireland's 1,500 filling stations had run dry. The army was deployed. The Garda declared an exceptional event. A plumber at one blockade told RTÉ he simply could not continue doing business at these fuel prices.

The Irish government had already cut excise by €250 million, but it was not enough and the pump swallowed the relief before anyone felt it.

Your AI prices are about to jump

A single AI query uses roughly a thousand times more electricity than a traditional web search. An AI server rack draws 40 to 100 kilowatts compared with five to 15 for a conventional one. Global data centre electricity consumption is projected to exceed 1,000 terawatt hours this year, equivalent to the entire annual consumption of Japan.

In March, Iranian drones struck Amazon Web Services facilities in the UAE and Bahrain, causing service disruptions across the region. For the first time in modern conflict, commercial data centres became explicit military targets. Iran's Revolutionary Guard has since threatened to destroy the gigawatt-scale Stargate data centre being built for OpenAI in the Emirates.

One academic stated that the scaling laws that drove the AI boom are "fundamentally peacetime constructs," built on the assumption that energy supply is infinitely elastic.

The war is repricing that bet.

Gulf states have committed roughly $300 billion to AI infrastructure.

That capital pipeline depends on oil export revenues now being choked by the very strait the war was supposed to liberate. The helium required for semiconductor fabrication transits the same corridor. The chips that power the servers that run the models that answer your questions trace back, eventually, to the same barrel of oil and the same waterway.

The top line: growth gets squeezed as capital that was earmarked for expansion gets sucked into paying for energy.

The bottom line: companies have never needed AI more, but if you are cutting headcount on the assumption that compute costs stay flat, think again. The pricing model you built in January is now tied to the Strait of Hormuz and that's not going to end soon.

So you might be a bit more thankful for your overhead on two legs than last month. Its pricing is steadier.

Happy Monday.

Dan

The Proposition is running free growth diagnostics for two companies this month. If something's holding you back and you can't see what it is, reply to this email.

Artificial intelligence (AI) is driving investments in infrastructure globally. The European Commission plans to invest $20 billion to build four "AI gigafactories" to enhance its AI capabilities. Meanwhile, developers CIM Group and Novva Data Centers have secured a $2 billion loan to construct a 100-acre AI data center in Utah. However, these massive projects raise concerns about environmental impacts, including significant energy consumption and electronic waste.

OpenAI has secured a $12 billion, five-year contract with cloud computing provider CoreWeave to supply the necessary computing power for training and operating its AI models. This partnership allows OpenAI to reduce its dependency on Microsoft for computing resources, with Microsoft agreeing to spend $10 billion on CoreWeave's services by 2030. CoreWeave, which transitioned from cryptocurrency mining to cloud services, has become a notable provider for AI development using Nvidia's GPUs.

The UK Atomic Energy Authority (UKAEA) and Eni are building the world's largest and most advanced tritium fuel cycle facility, set to open in 2028 at Culham Campus, Oxfordshire. The site will focus on recovering, storing and recycling tritium—a crucial step toward making fusion energy commercially viable.

Microsoft plans to invest an additional $297 million in artificial intelligence infrastructure in South Africa. The investment aims to enhance AI capabilities in the region and includes funding technical certification exams for 50,000 individuals to develop high-demand digital skills. This initiative aligns with Microsoft's broader strategy to expand data centers for training AI models and deploying AI and cloud-based applications.

Siemens Energy has entered into an exclusive agreement to supply conventional technology, including steam turbines and generators, for Rolls-Royce's small modular nuclear reactors (SMRs). These Generation 3+ modular nuclear power plants are designed to be more scalable, faster and cost-effective to deploy compared to traditional large-scale nuclear plants. The partnership is expected to boost job creation and manufacturing opportunities in the UK. Rolls-Royce's SMRs, each with up to 470 megawatts output, have been selected for projects in the Czech Republic and are under consideration in Britain and Sweden.

Last Energy has announced plans to construct 30 microreactors in Haskell County, Texas, to meet the increasing energy demands of the state's data centers. The company has initiated steps for development, including filing for grid connection with ERCOT and an "Early Site Permit" with the U.S. Nuclear Regulatory Commission. Data centers in Texas, developed by companies like OpenAI and Meta, consume significant amounts of power, prompting the exploration of nuclear power as a cleaner and more efficient energy source.

Equinix has forecasted its first-quarter and annual revenue to be below analysts' estimates, attributing this to the impacts of a strong dollar and significant spending on data center expansion, influenced by the growing adoption of generative AI. The company's extensive international presence makes it susceptible to foreign exchange rate fluctuations.

Researchers at The Ohio State University have developed a prototype battery that converts gamma radiation from spent nuclear fuel into electricity. Tests demonstrated the battery's potential, suggesting future applications in environments with high radiation levels, such as nuclear waste storage or space exploration.

Welcome to this month’s Nuclear News Round-Up.

The nuclear sector is charging forward, with small modular reactors (SMRs) securing major deals, fusion energy breaking records and the UK government doubling down on nuclear expansion.

Here’s a look at the biggest nuclear stories shaping the future of energy this month:

• Deep Fission Powers Up Data Centers with Mile-Deep SMRs Deep Fission, in partnership with Endeavour Energy, will supply 2GW of nuclear power to its global data center portfolio. Their borehole reactor technology, set to go online in 2029, promises safe, low-cost, zero-carbon power at just 5-7 cents per kWh. With major tech firms like AWS, Google and Switch investing in SMRs, nuclear energy is becoming an essential component of AI-driven infrastructure. Read more
• China’s ‘Artificial Sun’ Breaks Fusion Record China’s Experimental Advanced Superconducting Tokamak (EAST) has achieved a new milestone by sustaining fusion power for a record 1,066 seconds (almost 18 minutes). While still experimental, these advancements suggest a promising future for fusion energy. With new facilities underway in Anhui Province, China is clearly pushing forward its fusion energy agenda. Read more

• UK Government’s ‘Build Baby Build’ Nuclear Expansion Prime Minister Keir Starmer has announced plans to overhaul planning rules for SMRs, positioning the UK as a potential global leader in nuclear energy. Despite criticisms regarding cost and feasibility, this initiative is celebrated for its potential to enhance energy security, create jobs and support the UK’s net-zero objectives. Read more
• Rapid Fusion Unveils UK’s First Large-Format Hybrid 3D Printer Rapid Fusion’s new Medusa 3D printer is revolutionizing the production of critical nuclear components. Funded by £1.2 million from Innovate UK, this technology promises faster production, greater precision and reduced costs, essential for manufacturing SMR parts and fusion energy components. Industry leaders like Rolls-Royce are already leveraging this breakthrough. Read more
• Tokamak Energy Selects Aras PLM to Accelerate Fusion Development Fusion technology leader Tokamak Energy has implemented Aras Innovator SaaS to enhance their development processes. This platform will help manage complex fusion simulations and streamline the production of high-temperature superconducting magnets, boosting their commercial fusion efforts. Read more

See you next month

The Greenbang Team

Thanks to Giant Leap Digital and Crypto Marketing World for their support of Greenbang.

We're building a new segmentation model of the different types of data centres vs the role of nuclear power in them for a report we're writing.

While nuclear power offers advantages in reliability, sustainability and baseload power, factors like safety, waste disposal, public perception and regulatory hurdles must be addressed.

Implementing heat-sharing schemes requires careful planning, infrastructure development and collaboration with local communities and industries.

SMRs: Small Modular Reactors hold significant promise for increasing the flexibility and applicability of nuclear power to different data centre types, particularly enterprise and edge deployments.

GOOGLE AND KAIROS POWER PARTNER TO DEPLOY 500 MW OF CLEAN ELECTRICITY GENERATION
The deal represents the first corporate agreement for multiple deployments of a single advanced reactor design in the United States.

Alameda, CA – October 14, 2024 – Kairos Power and Google have signed a Master Plant Development Agreement, creating a path to deploy a U.S. fleet of advanced nuclear power projects totaling 500 MW by 2035.

Under the agreement, Kairos Power will develop, construct, and operate a series of advanced reactor plants and sell energy, ancillary services, and environmental attributes to Google under Power Purchase Agreements (PPAs). Plants will be sited in relevant service territories to supply clean electricity to Google data centers, with the first deployment by 2030 to support Google’s 24/7 carbon-free energy and net zero goals.

The innovative, multi-plant agreement will support technology development by extending Kairos Power’s iterative demonstration strategy through its first commercial deployments. Building on progress from the early iterations, each new plant will enable continued learning and optimization to support accelerated commercialization. Along the way, milestone-based accountability baked into the agreement will establish confidence in Kairos Power’s ability to deliver throughout the long-term partnership.

“Our partnership with Google will enable Kairos Power to quickly advance down the learning curve as we drive toward cost and schedule certainty for our commercial product,” said Mike Laufer, Kairos Power CEO and co-founder. “By coming alongside in the development phase, Google is more than just a customer. They are a partner who deeply understands our innovative approach and the potential it can deliver.”

“Having an agreement for multiple deployments is important to accelerate the commercialization of advanced nuclear energy by demonstrating the technical and market viability of a solution critical to decarbonizing power grids while delivering much-needed energy generation and capacity,” said Jeff Olson, Kairos Power Vice President, Business Development & Finance. “This early commitment from Google provides a strong customer demand signal, which reinforces Kairos Power’s continued investment in our iterative development approach and commercial production scale-up.”

Google’s deep commitment to decarbonization makes them a leader in clean energy development. Since 2010, the company has signed more than 115 agreements totaling over 14 GW of clean energy generation capacity. The additional generation that will be developed under this multi-plant agreement with Kairos Power will complement Google’s existing use of variable renewables, like solar and wind, and help them reach their ambitious 24/7 carbon-free energy and net zero goals.

“This landmark announcement will accelerate the transition to clean energy as Google and Kairos Power look to add 500 MW of new 24/7 carbon-free power to U.S. electricity grids,” said Michael Terrell, Google Senior Director of Energy and Climate. “This agreement is a key part of our effort to commercialize and scale the advanced energy technologies we need to reach our net zero and 24/7 carbon-free energy goals and ensure that more communities benefit from clean and affordable power in the future.”

Kairos Power applauds Google for its continued leadership as a first mover, helping to bring advanced reactor technology to market, as they have done with other clean energy technologies. Google’s commitment will catalyze new nuclear development, support U.S. decarbonization goals, and expand access to safe, clean, and affordable nuclear energy aligned with Kairos Power’s mission: To enable the world’s transition to clean energy with the ultimate goal of dramatically improving people’s quality of life while protecting the environment.

In the US alone, consumers spend around $350 billion annually on electricity, with three-quarters of this energy going to waste. As energy consumption rises, the inefficiencies in current systems are becoming a visible challenge.

The electrification of transportation, heating, and industrial processes will add a significant load to the grid. Similarly, the shift towards electric heating solutions, such as heat pumps, will increase electricity demand during peak heating periods.

Here are the seven areas I am watching:

1 - Small Modular Reactors (SMRs) represent a significant advancement in nuclear technology, offering a promising solution to meet future energy needs sustainably.

SMRs are more flexible and cost-effective than traditional nuclear reactors, making them suitable for various applications, including powering remote locations and integrating with renewable energy sources.

Rolls-Royce is leading efforts to develop SMRs, with plans to establish a testing facility at Sheffield University.

This initiative is part of a broader strategy to deploy SMRs worldwide, providing a stable, low-carbon energy source. In Poland, the Ministry of Climate and Environment has approved plans for Rolls-Royce SMRs, highlighting the growing international interest in this technology.

SMRs offer several advantages over traditional nuclear reactors. They are safer due to their smaller size and advanced safety features, which reduce the risk of catastrophic failure. Additionally, their modular design allows for faster construction and scalability, enabling utilities to add capacity as needed. The ability to generate reliable, low-carbon energy makes SMRs key in transitioning to a sustainable energy grid.

2 - Advances in energy storage technologies, such as batteries and pumped hydro storage, are improving the reliability and efficiency of renewable energy systems. For instance, Tesla's Megapack battery systems store excess solar energy and stabilise the grid.

3 - Smart-grid technologies enable real-time monitoring and management of electricity distribution, enhancing grid reliability and efficiency. These systems use sensors, communication networks, and data analytics to optimise energy flow, detect faults, and balance supply and demand. 

An example is the development of the Smart Grid Demonstration Project in Austin, Texas, which aims to integrate renewable energy, energy storage, and smart meters to create a more resilient grid.

4 - AI algorithms can predict energy demand patterns, optimise grid operations, and manage distributed energy resources. Google's DeepMind has used AI to improve the energy efficiency of its data centres, reducing cooling energy use by 15%.

5 - Existing grid infrastructure in many regions needs to be in the right place. Building and upgrading these systems requires significant investment and coordination.

6 - The regulatory landscape in the United States varies significantly by state, creating challenges for the nationwide deployment of advanced energy technologies.

7 - Cybersecurity:  The 2021 ransomware attack on Colonial Pipeline highlighted the vulnerability of critical infrastructure and underscored the need for enhanced cybersecurity protocols.

Great interview by Dan O’Shea with NetMind.Ai’s CEO, Kai Zou, on how the race to obtain compute power is forcing people to adopt decentralised compute models. 

Even investors are stockpiling their own GPUs now to ensure their startups have enough power when they need it. 

Decentralised compute is not new. It draws on the combined processing power of multiple machines across different machines in different locations. It’s commonly used in crypto mining and back when I were a lad, the SETI screen saver used the combined compute power of millions of machines to analyse radio signals for possible signs of extra terrsialt life.  

Octopus Energy Generation acquires 100% of Exagen Group from founder Jeremy Littman, driving forward UK expansion and advancing its renewable energy pipeline to support national climate goals.

Highlights:

• Exagen gains backing from new ownership to strengthen its position in the UK energy sector.
• The company’s development pipeline includes over 2.4GW of solar and energy storage projects across England.
• This follows Octopus’s initial investment for a 24% stake in August 2022.

24 September 2024: Exagen Group, founded by CEO Jeremy Littman, has finalised a deal for a 100% equity buyout by Octopus Energy Generation, one of Europe’s largest specialist renewables investors. This acquisition, on behalf of the Octopus Energy Development Partnership (OEDP) fund, managed by Octopus, is set to propel Exagen’s ambitious growth plans as it continues to contribute to the UK’s green energy transition.

The buyout follows Octopus’ initial investment in August 2022 for a 24% stake in the business. As Exagen’s project pipeline progresses toward construction, the new ownership will support its long-term growth. Octopus Energy Generation currently manages nearly £7 billion of renewable energy projects across the globe.

In the past year, Exagen has secured planning consent for 520MW of standalone battery storage projects at sites in Leicestershire and Gloucestershire. Additionally, the company has 260MW of solar and 268MW of battery storage either in planning or poised to enter planning in the coming year. Exagen is committed to integrating solar and energy storage infrastructure with natural habitats to enhance biodiversity, improve soil health, and provide community amenities by the end of each project’s lifecycle.

 Jeremy Littman, CEO of Exagen Group, commented: “I founded Exagen to deliver and manage a robust pipeline of UK renewable energy projects with the mission of fighting climate change. Having built a talented team to drive this vision forward at pace, I am proud to pass the baton to Octopus , who will continue to deliver projects and expand the platform to new heights.”

“The initial investment from Octopus provided the foundation for Exagen’s success. With this buyout, Octopus is now positioned to further advance our mission of meeting the UK’s renewable energy targets by 2035.”

 Alex Brierley, Co-head of Octopus Energy Generation’s fund management team, commented: “We first invested in Exagen two years ago and we’ve been impressed by their strong pipeline of solar and storage projects in development that will help boost British energy security. By increasing our investment, we’re going to turbocharge the build-out of new green energy and rapidly grow the UK’s energy storage capacity.”

The rapid expansion of artificial intelligence (AI) is revolutionizing industries, but it also brings significant energy demands, particularly in the tech sector.

While AI can enhance energy efficiency in some areas, it also contributes to rising greenhouse gas (GHG) emissions due to its immense computational requirements. This has led to a growing debate about whether AI can genuinely support sustainability efforts or whether it will exacerbate environmental challenges.

A primary concern is the energy required to run AI models, especially the training of large models like GPT-3 and GPT-4. Training GPT-3 consumed about 1,300 megawatt-hours (MWh) of electricity, roughly equivalent to the annual consumption of 130 U.S. households. GPT-4, an even more advanced model, required up to 50 times more electricity​ (World Economic Forum)​(Forbes ME).

These models are just the beginning. As AI grows, the demand for data centers, which host and run these models, is soaring. In 2023, Google reported a 48% increase in its GHG emissions compared to 2019, driven primarily by AI-related data center energy consumption. Microsoft has experienced similar trends, with its emissions rising nearly 30% from 2020​ (Forbes ME)​(Brown Advisory).

Despite these energy challenges, AI has the potential to optimize energy use in other areas. For instance, Google’s application of machine learning reduced energy used for cooling its data centers by 40%. AI's ability to predict energy demands and manage grid efficiency is helping to integrate renewable energy sources more effectively​ (World Economic Forum)​(Brown Advisory).

However, the energy consumption of AI models is expected to increase as new models and applications emerge. Tech giants like NVIDIA are addressing this by developing more energy-efficient hardware. NVIDIA’s latest chips offer a 30x performance improvement while using 25x less energy, providing a glimpse of how technology might mitigate some of AI's environmental impact​ (World Economic Forum).

Artificial intelligence (AI) is advancing rapidly, with its applications influencing industries worldwide. This surge is particularly evident in the growing adoption of generative models like ChatGPT, GPT-4, and other large language models. However, as AI becomes more entrenched in daily operations, the energy demands associated with training and running these models are soaring, putting immense pressure on global data centers.

The primary challenge for data centers is to manage the energy-intensive demands of AI while maintaining operational efficiency. These data centers, crucial for AI model training and storage, are now struggling to balance AI's expansion with sustainable energy practices. The environmental implications of this trend are significant, as AI-driven data centers account for a rising percentage of global energy use, leading many to question the long-term sustainability of AI.

The Rising Energy Crisis in AI

Training large AI models requires massive computational resources. Models like GPT-3 consumed about 1,300 megawatt-hours of electricity during training, enough to power 130 average U.S. households for a year​(

World Economic Forum). As newer models such as GPT-4 become more sophisticated, their energy requirements multiply exponentially. AI compute demand is now doubling every 100 days, a statistic that illustrates the sheer scale of the energy crisis unfolding in the AI industry​(World Economic Forum).

AI systems are not just data-heavy; they are also "always-on" systems. For instance, models like ChatGPT run continuously to respond to user queries, placing additional stress on the energy grid. Data centers, which already consume approximately 2-3% of global electricity, are experiencing even greater strain due to the rising demands of AI-powered services​ (LL MIT).

Challenges for Data Centers

Data centers face two major challenges: scaling up their computational infrastructure to handle AI's growing workloads and addressing the corresponding surge in energy consumption. The infrastructure necessary to support large-scale AI deployments requires vast amounts of power, primarily because of the heavy reliance on energy-intensive hardware such as GPUs (Graphics Processing Units). For instance, Nvidia's hardware innovations are designed to improve AI performance, but they also come with significant energy demands ​(BigDATAwire).

Moreover, AI workloads, particularly those involving generative models, differ from traditional data center tasks in both scale and complexity. As a result, data centers must rethink their energy strategies. Increasing the number of servers or GPUs is only part of the solution. To truly address the energy crisis, they must also innovate in areas such as energy-efficient cooling and smarter workload scheduling.

Pathways to Energy Efficiency

Despite the challenges, there are promising approaches being deployed to mitigate AI's energy consumption. Power capping is one strategy where the power drawn by GPUs during AI training is limited. This technique can reduce energy use by around 12-15%, though it increases training times slightly​ (LL MIT). While the impact on speed is minor, the energy savings can be substantial over time.

Another development is the growing use of advanced hardware that improves energy efficiency without sacrificing performance. Nvidia's latest GPUs, for example, offer a 5x improvement in energy efficiency compared to older CPU-only systems​(

BigDATAwire). These GPUs are helping reduce the overall energy consumption of AI tasks, making large-scale deployments more sustainable.

Cooling technologies also play a crucial role. Data centers are exploring innovative ways to minimize the energy required for cooling, such as running workloads at night when temperatures are lower or employing liquid cooling systems. These methods not only reduce the environmental impact but also extend the lifespan of data center hardware by preventing overheating​ (LL MIT).

Looking Forward: Sustainable AI

The road to sustainable AI will require a multi-faceted approach, combining technological advancements with strategic operational changes. Data centers need to embrace renewable energy sources, such as solar and wind, to reduce their carbon footprint. Additionally, more efficient data management practices, like minimizing the storage of "dark data" (unused data), can lead to significant energy savings​ (World Economic Forum)​(LL MIT).

Regulatory measures will also play a role in driving sustainability. Governments and international organizations are beginning to implement stricter energy efficiency standards for AI deployments, requiring transparency in reporting energy usage and emissions​ (Nature). This could help push AI developers and data centers toward more energy-conscious practices.

As AI continues to evolve and integrate into everyday systems, its energy footprint will become an increasingly pressing issue. By focusing on both innovation and sustainability, the AI industry can work toward balancing technological growth with environmental responsibility.

Hofstra University's DeMatteis School announced that it received a $1 million grant from the National Science Foundation (NSF) to enhance its high-performance computing capabilities.

This funding will support the development of new computational resources and training programs aimed at fostering research and innovation in various scientific fields. The grant underscores the importance of HPC in advancing academic research and preparing students for careers in technology and science.

In the face of climate change and the urgent need for sustainable energy solutions, modular nuclear reactors (MNRs) are gaining significant attention as a viable alternative to traditional nuclear power plants. These small modular reactors (SMRs) promise enhanced safety, reduced costs, and greater flexibility, making them an attractive option for countries looking to diversify their energy portfolios.

Understanding Modular Nuclear Reactors

Modular nuclear reactors are designed to be built in factories and transported to sites for assembly, allowing for a more streamlined construction process compared to conventional nuclear plants. Their smaller size and modular design enable utilities to scale capacity as needed, which can be particularly beneficial for regions with fluctuating energy demands. Additionally, SMRs are typically designed with advanced safety features, reducing the risk of accidents and enhancing public confidence in nuclear energy.

The Global Landscape of Modular Nuclear Reactors

The rise of SMRs is not just a trend but a response to global energy challenges. Countries around the world are exploring this technology as a means to achieve energy security while reducing greenhouse gas emissions. According to the International Atomic Energy Agency (IAEA), the development of SMRs is being closely monitored, with numerous projects underway across various nations, including the United States, Canada, and several European countries.

There are several key players in the Modular Reactor Market

1. NuScale Power: Based in the United States, NuScale is a frontrunner in the SMR market. Its design features a 60-megawatt reactor that can be scaled up by adding modules, making it suitable for a range of applications. The company has secured significant funding and is working towards regulatory approvals for its first plant in Idaho.
2. Westinghouse Electric Company: A historic player in the nuclear industry, Westinghouse is developing its own SMR technology, known as the eVinci micro-reactor. This design focuses on providing power in remote locations and is expected to play a crucial role in future energy solutions.
3. GE-Hitachi Nuclear Energy: This joint venture is making strides with its BWRX-300, a 300-megawatt SMR that leverages existing boiling water reactor technology. The BWRX-300 aims to significantly lower construction costs and improve the efficiency of nuclear power generation.
4. Rolls-Royce: In the UK, Rolls-Royce is advancing its own SMR initiative, which aims to produce reactors capable of generating 470 megawatts. The company emphasizes the economic benefits of its design, projecting that it could provide low-cost energy while contributing to the UK’s net-zero goals.
5. TerraPower: Founded by Bill Gates, TerraPower is developing the Natrium reactor, which combines a sodium-cooled fast reactor with a molten salt energy storage system. This innovative approach allows for flexible energy generation, aligning with the variable nature of renewable sources like wind and solar.

The Future of Modular Nuclear Reactors

As nations strive to meet their climate commitments, the modular nuclear reactor market is poised for substantial growth. The combination of advanced technology, regulatory support, and increasing investment from both public and private sectors indicates a promising future for SMRs. Moreover, the ability of these reactors to complement renewable energy sources positions them as a critical component of a sustainable energy strategy.

References

1 - World Nuclear Association. (2024). Small Nuclear Power Reactors. Read more 

2 - TVA Newsroom. (2024). TVA Board Approves Additional $150 Million in Advanced Nuclear Funding. Read more 

3 - Enerdata. (2024). Small Modular Reactors: Advancing Nuclear Power Generation. Read more 

4 - IAEA. (2023). What are Small Modular Reactors (SMRs)? Read more 

The notorious Three Mile Island nuclear facility in Pennsylvania is poised to restart operations after a five-year hiatus.

This move comes as part of a strategic agreement between Constellation Energy, the plant's operator, and Microsoft, which seeks to power its rapidly expanding artificial intelligence initiatives.

Three Mile Island is historically significant for experiencing the worst nuclear incident in U.S. history in March 1979, when a malfunctioning valve led to a coolant loss and overheating in the Unit 2 reactor. While Unit 2 remains decommissioned, the nearby Unit 1 was shut down in 2019 due to financial challenges. Now, it is set to resume generating approximately 835 megawatts of electricity for the Pennsylvania energy grid.

This event marks a historic first, as it will be the initial instance of a U.S. nuclear reactor being reactivated after closure. The initiative is anticipated to generate around 3,400 jobs and inject at least $16 billion into the local economy. In a nod to this new chapter, the facility will be renamed the Crane Clean Energy Center, honoring Chris Crane, the former CEO of Constellation’s parent organization.

JAs energy demands escalate, particularly from major tech players like Microsoft, Google, Amazon, Meta, and Apple, environmental concerns are becoming increasingly pressing.

NVIDIA has made significant strides in 2023 with its energy-efficient AI computing platforms, introducing new chips and systems designed to drastically reduce power consumption while maintaining top-tier performance. The company's Grace Hopper GH200 Superchip has been a standout in this space, leading the charge in energy-efficient AI and high-performance computing (HPC).

NVIDIA's GH200 Superchip delivers over 1,000x more energy efficiency than previous-generation GPUs, particularly in tasks requiring mixed precision and AI workloads. This breakthrough architecture has been adopted by supercomputing centers worldwide, including Germany's JUPITER system, where the superchip achieved 72.73 GFlops per watt, making it one of the most energy-efficient supercomputers in operation today. Such systems are essential for AI applications, from climate science to advanced medical research, enabling faster results while consuming less energy​ .

Additionally, the National Energy Research Scientific Computing Center (NERSC) found that using NVIDIA’s A100 Tensor Core GPUs resulted in a 5x increase in energy efficiency compared to CPU-only systems. This improvement is critical as data centers look to reduce their carbon footprints amid rising AI compute demands​

NVIDIA's ongoing focus on energy efficiency is helping redefine the AI industry, ensuring that the growth in AI-powered applications doesn't come at an unsustainable environmental cost. With systems like Grace Hopper setting new standards for efficiency, NVIDIA is paving the way for a greener future in AI and supercomputing​.

The University of Alabama announced the opening of a new high-performance computing center aimed at enhancing innovation and discovery. This facility is expected to provide advanced computational resources to support research across various disciplines, enriching student experiences and fostering collaboration among faculty. The center will play a critical role in driving research initiatives and supporting projects that require significant computational power, thus positioning the university as a leader in HPC capabilities. Read more here