Up until recently nuclear has had a simple offer for electricity buyers. It provides a huge amount of power from a plant that requires little land at the point of dispatch. It does it consistently, 24 hours a day and for decades. What is more, this century, its status as a low carbon generator has made it more attractive. But now its customer proposition has to change.
The offer of so-called ‘baseload’ power may be a compelling proposition if you have a several large buyers, or a central buyer, that have a need for large, steady power supply. Now the numbers of such customers are shrinking and the way they interact with the electricity industry is changing too.
An example of nuclear’s ideal customers might be a group of industrial users who require power for decades and will not have the option to relocate. But even those users may change the way they use power over the lifetime of a nuclear plant, which in current discourse may reach a century. Of course some customers will continue have a need a need for bulk power and they will no doubt be willing to sign contracts, some lasting several years, with inflexible generators. But they are unlikely to contracts for more than a few years, because they see the risk and benefit changing and they do not want to lose the ability to change their approach. Recently, industrial customers have looked at their own potential energy resources – whether that is a waste product that can feed a thermal power plant, or the roof and land area that allow for on-site renewables – and used them to reduce the power needed from the market. In that case they now need variable power imports that can respond to the rise and fall of self-generated power or act as an emergency backup.
Other typical nuclear customers are national monopoly electricity suppliers. In practice, this group of nuclear customers always had to make sure nuclear was part of a portfolio. Each electricity supplier had end-customers to serve who generally need variable supply to match their usage over days, weeks and months, not unchanging baseload. It was suppliers who took on that job – and the financial risk of being ‘out of balance’ between supply and demand – combining nuclear with complementary sources to provide the variable supply needed by their customers.
The outdated baseload concept
It was the case that the price of nuclear-generated power was fairly predictable – if not always the cheapest option. It presents a different risk profile compared with thermal plants, which are relatively cheap to build but whose actual ‘cost to generate’ depends heavily on the fuel price. The link with fuel costs has dramatically altered the use of thermal plant in both long and short term. For example, the discovery of gas resources in the 1980s and 1990s meant it became a fuel of choice and we are still living with the switch to gas turbines that resulted – and the consequences of the dramatic changes in gas supply that followed Russia’s invasion of Ukraine.
That recent experience is helping make the risk profile of high capital cost, low fuel cost, plant like nuclear more attractive. But it makes renewable energy sources even more attractive, because they also offer low running costs but their build profile is one of relatively small units installed quickly, so that even multi-gigawatt scale wind farms can be built in phases that allow generation of power – and revenue – within a few months.
The changing power mix means nuclear can no longer rely on the old concept of baseload power. In fact 2025 will mark a decade since Steve Holliday, then the chief executive of GB’s system operator National Grid, dismissed the idea of baseload in an interview for the World Energy Council. Back in 2015 he said “this is an industry that was based on meeting demand. An extraordinary amount of capital was tied up for an unusual set of circumstances: to ensure supply at any moment.”
Holliday was speaking at an earlier stage of the rollout of renewables. Since then the renewable rollout has increased by orders of magnitude. Offshore wind turbines have stepped up in size from 3 MW to 16 MW and are still growing, and offshore wind arrays are typically planned in the single or multi gigawatt scale.
But Holliday also pointed out the potential at small and local scales: “The idea of baseload power is already outdated. I think you should look at this the other way around. From a consumer’s point of view, baseload is what I am producing myself. The solar on my rooftop, my heat pump – that’s the baseload. Those are the electrons that are free at the margin.” Holliday’s prediction has come to pass, as the cost of solar PV has continued to tumble (see box) and deploying it has become familiar for both domestic and business customers. Some scenarios see ‘PV everywhere’ in future. That will alter long-established profiles for business and domestic use as ‘behind the meter’ PV meets some or all of the user’s needs, but it will also add to the volatility to be managed as generation rises and falls with daylight hours and with local weather.
Ancillary services and nuclear
Nuclear has to prepare to lose its ‘baseload’ customers and consider how it can satisfy different customer needs in a very different power industry. Who will be nuclear’s future customers and how can it meet their needs?
Clearly flexibility will be at a premium. But Holliday also said, “I believe there will be different answers for different places, rural and cities, and for different customers.”
That tends to suggest that the industry’s move towards small modular reactors (SMRs) is the right one, and that is only partly because of the need for flexibility in power level. SMRs also allow for geographical flexibility.
System operators need flexibility – upwards as well as downwards. The electricity network is built around the characteristics of thermal power plant, and without such plant on the system it is harder to maintain system stability. Thermal plants’ heavy rotating generators are useful in provide physical ‘inertia’, which tends to pull the supply back towards its defined operating frequency. They can help manage voltage levels, by either absorbing ‘reactive power’ or producing it. Sometimes, in a system that has high levels of renewables, on a day with sun and wind renewables can supply most of the energy required by users. As a result the system operator may ask users to increase their power demand, so it can bring thermal plant into action to provide those services – alternatively it has to constrain off renewables, often the cheapest plant on the system, to bring the necessary thermal plant online.
In the case of voltage support, the requirement is location-specific so it requires distributed generating sources (or other types of flexibility) and because it is determined partly by the types of local demand and whether they supply or absorb reactive power, it may change over time.
All these specific requirements should tend to benefit SMRs, because they have more flexibility over siting and a shorter lifetime.
Nuclear has to be willing and able to provide these services, as well as flexibility for the system operator or to meet large customers’ needs around their self generation. That flexibility has proved problematic in the past, not so much because of technical incapacity but because the economic balance has not been in its favour. Currently, a flexibility payment has to be substantial and longstanding if it is to make up for lost generation, increased stresses on plant systems and – unique to the nuclear industry – the cost and time involved in developing a safety case to ensure the regulator is comfortable with such an operating regime.
Nuclear as part of a system
One answer is for nuclear suppliers to operate their nuclear capacity not as a standalone entity but as part of a system. Michael Liebreich makes this point when he says the answer to variability is not any single technology, although batteries have been a focus of investment. Instead, “it is a system solution – a combination of demand response, interconnections, excess generating capacity, pumped storage, nuclear power, CCS, hydrogen and biogas long duration storage, integrated by means of an extensive grid and managed using the latest digital technologies”.
This is already happening elsewhere. In the UK now storage in the form of distributed batteries, some standalone and some in applications as small as electric vehicles, can participate in contracts to supply the system operator with capacity or flexibility as part of a ‘virtual power plant’ aggregating tens, hundreds or thousands of units. Elsewhere, large-capacity batteries are increasingly being co-located at wind farm or solar farm sites. They allow efficient use of the grid connection, but also they give the operator the tools to offer flexibility or fixed supply, as required, to customers.
The new watchword is ‘whole system’ management of flexibility: using excess power in a variety of ways. In downward flexibility whether that is to charge vehicle batteries, to boost large heat stores in district heat systems, to electrolyse hydrogen or to prime other forms of storage such as pumped hydro or compressed air. In some cases a nuclear plant may be consistently allocating part of its production for such customers, with an agreement to reduce supply where extra power is needed by the system operator; in other cases the customer may benefit from accepting extra energy because the system operator has asked the nuclear operator to reduce exports to the grid. Both, of course, depend on the type of flexibility available. But an active and responsive generator with a portfolio
of sources can use fast and slow flexibility in concert, and allowing, for example, a battery to respond within a few seconds, with nuclear load following more slowly and other sources filling the gap.
The need for flexibility has driven the industry towards small modular reactors. But with the right attitude it is possible that large nuclear too can retain a place in the power portfolio. Nuclear’s bulk supply has had its disadvantages as well as advantages, requiring not only enough demand, but enough reserve and response in case of plant shutdown. But if, as envisaged, the future is one of ‘electrify everything’ including transport and heat, electricity demand multiplies and the drive to large scale installations applies across the electricity industry.
Alongside gigawatt-scale nuclear will be similar scale offshore wind farms and interconnectors. Great Britain is a case in point: at 1250 MW Sizewell B was for many years the largest in-feed in the country’s grid. But the East Coast, where it is situated, is also now the landing site for the world’s largest offshore wind farm, Dogger Bank, which will total 3.6 GW when complete. The region also has two gigawatt-sized interconnectors, which could see a shift of 4 GW in short order if they both switched from full import to full export. Large nuclear may no longer be the grid’s largest potential in-feed risk.
The question for nuclear operators is how they can use nuclear’s strengths to meet the new needs of customers who are no longer seeking long-term contracts for bulk power. If nuclear alone cannot respond in a way that meets market needs, how can operators find partnerships where the whole is greater than the sum of the parts and customers get a compelling proposition? Nuclear can’t be a large standalone supplier any more. It has to be part of a more responsive package.
Will solar eat everyone’s lunch?
A major energy story of the 2020s has been the immense rollout of solar – from utility scale to household panels.
In a February blog post Michael Liebreich, senior contributor at Bloomberg New Energy Finance named solar power as a ‘superhero’ of the drive to Net Zero.
He compared 2004, which saw 1 GW of solar PV installed with 2023, when there were several occasions when 1 GW of solar PV was installed in a day.
He said that with last year wind and solar added enough capacity to deliver 3,400 TWh, or 3% of global power demand – and the rate of deployment is accelerating.
Solar is being deployed at utility scale. In 2023 the largest solar plants worldwide were in India and China. The 2.7 GW Bhadla Solar Park in the Thar Desert of Rajasthan has an area of 56 square kilometres. But the Golmud Solar Park in China is already bigger – 2.8 GW – and China plans to expand it to 16 GW by 2030.
According to data from the Federal Energy Regulatory Commission (FERC), solar provided 49.3% of new generating capacity in the United States, 50% more than the year before. In total, solar added 18,356 MW to the grid in 2023, while natural gas added 11,024 MW and wind added 6,356 MW. DNV says that Boosted by IRA, solar and wind power will grow 15- and 8-fold respectively by 2050. Solar will become the largest producer of electricity by the mid-2030s, supported by favourable economics and enhanced policy support in the region. Wood Mackenzie agrees: in a report on South America it said that with advances in batteries and lower costs, solar will outpace wind growth rates and become the leading technology for expansion over the next ten years in Argentina, Brazil, and Chile where it expects 48 GW of new utility and distributed projects by 2033.
By 2050, solar will account for almost half of all electricity generated in North America.
But solar’s benefit is that it can be deployed at any scale. Recently BNEF has increased its solar forecasts in the US by 12 GW for the 2023-2030 period, compared with its previous outlook. It believes most of the extra 12 GW will be in home solar.
PV is already the second largest part of Brazil’s electricity capacity mix. Hydroelectricity still has 40% of the generation market with almost 110 GW installed, but in 2023, but PV uptake in Brazil grew at a rate of more than 1 GW per month and the cumulative installed PV capacity reached over 37 GW. The deployment rate is 60 W per person per year and is fast enough to double the installed capacity every two years.
What is more, 70% of the 1 GW installed in 2023 was rooftop PV, so most of it never entered the electricity network. There are 93 million rooftops in the country, and so far, fewer than 2.5% of them have a rooftop PV system installed.
