Many people who are generally not nuclear followers were intrigued last week by announcements from Google and Amazon that they would buy their energy from small modular reactors (SMRs).
Google has agreed to purchase up to 500 MW of electricity from multiple SMR units, which will be designed, built and operated by the California-based firm Kairos Power.
Amazon has an even wider plan which includes investing directly in X-Energy, an SMR developer, including taking two seats on its board, and teaming up with two utilities, Energy Northwest (Washington State) and Dominion (Virginia), to investigate the deployment of X-Energy’s SMRs.
Clearly, there is a lot of media and public interest, but many comments about this story lacked accuracy and some of the salient points. I figured that followers of Prospect deserved better, so I have written this piece to give you a short overview of what this is all about.
Why is Big Tech Interested in SMRs?
The driving factor for big tech’s interest in nuclear power is that Google, Amazon, and other datacentre operators need increasing amounts of reliable, abundant, and diversified clean energy to power their datacentres and meet their net-zero targets. The International Energy Agency estimates that the demand for datacentre power could rise from 460 TWh in 2022 to over 1000 TWh by 2026. This need is continuing to accelerate through the greater use of AI applications.
Google and Amazon have chosen to add nuclear power to their shopping list owing to its high energy density and always-on power supply, which simply can’t be matched by any other green energy sources. Nuclear proponents are delighted by big tech’s new-found enthusiasm, while sceptics continue to point out the ‘traditional’ nuclear sector concerns of cost, delay, and waste management.
Interestingly, both big tech companies have opted for advanced – and not yet produced – ‘Generation IV‘ reactor designs, which offer designs with significant advantages over most nuclear reactors and counter some of these concerns.
To explain this properly, we need a short dive into some physics and engineering.
Pressurised Water Reactors (PWRs)
The dominant type of nuclear reactor is the pressurised water reactor (PWR), which uses water as the coolant to transfer heat at around 250°C from the uranium fuel in the reactor core to a steam generator, which drives the turbines to produce electricity.
Challenges of PWRs: Complexity, Cost, and Safety
The whole system is kept under pressure to keep the water from boiling, which means the reactor vessel containing the core and accompanying systems must be made of thick, high-quality steel. Losing the water coolant to a leak is bad, as it uncovers the fuel and stops it from cooling down, meaning in the worst case, the fuel rods might fracture or melt with the release of the very radioactive material within.
Therefore, PWRs have a number of safety systems to reduce the risk of any fuel damage, and all of this is enclosed in a large containment building.
This, in turn, makes a PWR complex and expensive; perhaps not the optimal design, but it had vast sums invested during the Cold War to make it work in submarines and aircraft carriers, so the civil nuclear power industry leveraged that investment despite its compromises.
As a result, over the seventy years of civil nuclear power, the PWR design has become very well-understood and the dominant choice for providing power to a nation’s electricity grid, with some 310 PWR plants around the world or around 70% of the total.
To mitigate the cost and complexity, PWRs have grown significantly in size to maximise economies of scale, from around 200 MWe in the 1950s to today with, for example, EDF’s 1600 MWe EPR model at Hinkley Point.
However, their very scale and complexity mean that only governments can support the huge cost and risk of such megaprojects. Recent nuclear megaprojects have not come easily, with well-reported cost and time overruns.
Are Small Modular Reactors (SMRs) the Future of Nuclear Power?
Enter the concept of the small modular reactor (SMR), so-called because SMRs are smaller (usually less than 300 MWe, although the Roll-Royce SMR is 470 MWe) and can be manufactured off-site and assembled in modules. The intention is that the economy of mass production offsets the loss of economy of scale, making them attractive to a wider range of users who can fund the smaller capital cost more efficiently.
The principle of reliable cost reduction through mass production of SMRs has not yet been proven, but there is plenty of interest globally in the concept, with over 90 designs on paper and some individual projects already taking shape, such as the GE-Hitachi SMR being built at Darlington, Ontario. Whilst nobody has a firm price for a mass-manufactured SMR yet, the hope of the reactor designers is that the combination of mass manufacture and more digestible financing per unit allows the concept to have an attractive and commercially viable whole-life cost.
Comparing SMRs to Traditional PWRs
Many of these SMR designs, particularly those nearest to commercial readiness, are just smaller PWRs, such as three of the four types currently competing in the UK government’s SMR selection process (the fourth is GE-Hitachi’s SMR, which does allow the water to boil – maybe we can discuss the nuances of this in another article if it wins the competition). Despite the same historical origins, PWR SMR designs currently being developed have little in common with naval PWR nuclear power plants, which mostly use highly enriched fuel and have very different user requirements, so they can’t be repurposed.
However, other SMR designs exist that do not use water as the coolant, known as ‘advanced’ or ‘Generation IV’ designs. For example, the Kairos plant chosen by Google uses molten fluoride salt, Amazon’s X-Energy Xe-100 uses helium gas, and others use molten sodium or molten lead.
Advantages of Non-Water-Cooled SMRs
Using something other than water as the coolant confers several potential advantages:
- Firstly, the different properties of the coolant mean that the reactor can use more advanced fuel, such as ceramic-coated uranium pellets with a melting point of over 1800° C. These will not melt or breach under any foreseeable circumstance, implying more safety in the design.
- Secondly, the operating pressure required is much less, reducing the risk of a leak, requiring less size and complexity of the system components, and so implying more safety and less cost. This also means it is possible with some designs, such as the Kairos SMR, to refuel the reactor while it is running (no need to depressurise and open it up), reducing downtime and improving efficiency.
- Thirdly, the coolant’s ability to operate at higher temperatures with better heat transfer characteristics enhances the system’s efficiency and cost-effectiveness. It also opens up the possibility of providing direct heat for hard-to-abate industrial processes, such as steel and ammonia production, making it a financially attractive option.
- Fourthly, the combination of the fuel and coolant makes the reactor potentially very responsive. Its power can be varied quickly, which can be useful if it must respond to a rapidly varying demand, something that is not necessarily so easy with a water-cooled design. This is vital if you want to create a fully reliable and consistent power supply on the same demand grid as other renewables, which vary with the time of day and the weather.
- Fifth, and perhaps ultimately the driving factor, the increased safety profile potentially makes it much easier to convince the nuclear regulator that the plant’s siting and operation can be under less onerous conditions, significantly reducing cost and allowing many more options around how the plant could be used.
All these promised characteristics – cheaper, simpler, safer, smaller, flexible and easier to site – make these advanced SMRs very attractive to private energy-hungry clients who are comfortable with innovative technology, such as big tech companies.
Much of last week’s commentary focused on the yet unproven nature of these designs, noting that they have yet to be demonstrated in a commercial environment, and indeed, their whole premise relies on mass production at scale. But these ‘advanced’ designs are well understood. X-Energy’s Xe-100 is based on decades of experience from similar designs operated in the US, South Africa and Germany, and all the UK’s civil power reactors since 1956, bar one, have used gas coolant. Molten salt coolant, although sounding exotic, has been around since the 50s as well. The current round of innovation for these designs comes from the prospect of new clean energy use cases – such as datacentres – justifying the predominantly privately-led investment in materials, fuel and routes to market.
Handling Waste in Advanced SMRs
What about waste?
Most countries, including the US, where these projects are being developed, already require the plant owner to set aside funding for managing waste, decommissioning, and fuel spent, which will apply to SMRs. Used ceramic fuel pellets have been stored in the US for many years with no issue, although their robust nature makes them more problematic to disassemble. The long-term proposal in most countries for current high-level waste and spent fuel is a geological disposal facility, about which Prospect has written previously. Some advanced SMR designs – although not Kairos or X-Energy – are ‘fuel burners’, meaning that there is less spent fuel produced, and what is left is only radioactive for a few hundred years compared to many millennia for current PWR fuel, potentially helping with the issue of long-term storage.
What about the cost?
Large new nuclear plants, mainly owing to the large financing and regulatory costs, struggle to be cost-competitive with renewables in current energy markets. Some of the SMR designers, with an eye to market capture, are very bullish about their projected costs, citing the cost advantages outlined above, but that is a question that any SMR manufacturer can’t truly answer until they have built a good number of units to demonstrate those manufacturing economies.
Meeting the Needs of Big Tech
Comparing the prices of different SMR designs, large plants, and other renewable energy sources is increasingly missing the point.
The big tech firms seek a power source with particular characteristics to suit their own datacentre operating models and are willing to pay a certain amount for reliable, always-on power. The choice of nuclear, and other clean energy technologies, is maturing to provide a range of energy solutions, which means that directly comparing $/kWh is becoming less meaningful.
An Aston Martin and a Skoda are both automobiles and do a great job of taking you from A to B, but to compare their running costs is moot, as they are made for different use cases and different price points.
The Future of SMRs and Market Demand
SMRs have often earned the soubriquet ‘Smart Marketing Reactor’, and the hype around them was only exacerbated by last week’s announcements.
Nevertheless, the definitive entry of big tech firms into the game is noteworthy. It provides a demand signal to the nuclear sector and the wider energy market that they are convinced of the benefits of advanced SMR designs and encourages the scaling up of production.
There is plenty of other activity around SMR development deployment, which suggests that this is not just hype but a real push towards commercial reality, and I look forward to penning more articles about this for Prospect’s loyal followers.
I hope you found this article useful. I tried to make the physics and engineering concepts concise.
At Prospect, we help clients make sense of nuclear energy matters across the advisory spectrum. Do contact me to find out more.
John Warden
John Warden brings 35 years of experience in the nuclear and defence sectors to Prospect. He specialises in nuclear reactor project structure and financing, implementation of nuclear technologies, and strategies to meet climate goals using nuclear power. He is increasingly active in the field of advanced nuclear technology, where he advises on the economics and feasibility of deploying small modular reactors and advanced nuclear technology.
Contact John to arrange a call.
