Why SMRs and Solar Are Allies, Not Enemies
The clean energy debate keeps picking fights between technologies that actually need each other.
There is a argument that shows up constantly in energy forums, op-eds, and social media threads, and it goes something like this: nuclear is too slow and expensive for the climate crisis, solar is fast and cheap, therefore nuclear is obsolete. The counter-argument lands just as hard: solar is intermittent and can’t power civilization at night, therefore nuclear is necessary. Both sides present their case with the confidence of people who have never had to manage a power grid.
Neither is right. Or, more precisely, neither is complete. The choice between small modular reactors and solar is not a choice — it’s a false conflict dressed up as an engineering debate. The electricity system that most countries actually need in 2025 and beyond probably needs both. Badly. And the sooner advocates on either side figure that out, the sooner the clean energy transition actually accelerates.
The grid’s real problem is timing, not generation
Here is the core issue. Solar is spectacular at making cheap electricity when the sun shines. According to the International Energy Agency, solar PV generation grew 31% in the first half of 2025 alone — the fastest absolute growth on record. The Solar Energy Industries Association reports that solar accounted for 54% of all new electricity-generating capacity added in the United States in 2025. These numbers are genuinely astonishing. ☀️
But solar has a scheduling problem. The sun sets every evening around the same time that households turn on their stoves, their televisions, their lights. Electricity demand doesn’t politely follow a solar curve. What grid operators call the “duck curve” — the phenomenon where midday solar floods the grid, prices crash or go negative, and then demand surges at dusk just as solar drops to zero — is one of the central operational challenges of the 2020s energy system.
The challenge looks like this:
Solar generates maximum power roughly between 10am and 3pm, when demand is moderate
Evening demand peaks between 5pm and 9pm, when solar has largely vanished
The sharper the ramp between those two points, the harder it is for dispatchable generation to fill the gap
As more solar gets added, the midday trough gets deeper, the evening ramp gets steeper
According to TotalEnergies’ Energy Outlook 2025, the true cost of reliable 24/7 solar-based power, including storage, runs to about $100 per megawatt-hour — not the $30/MWh LCOE figure often cited for standalone panels in sunny regions. The IEA puts advanced nuclear, including SMRs, at around $63/MWh. At those numbers, nuclear stops being a curiosity and starts being competitive precisely in the hours and seasons where solar stumbles. 🔬
This isn’t an argument against solar. Solar is winning the daytime electricity market and should keep winning it. The question is what powers the grid during the other hours — and that question is where SMRs walk in.
What SMRs actually do when the sun shines
The old picture of a nuclear plant is a machine running flat-out, all hours, all year, indifferent to what the rest of the grid is doing. That picture is increasingly outdated, especially for SMRs.
Modern SMR designs — including GE Hitachi’s BWRX-300, NuScale’s design, and X-energy’s Xe-100 — are specifically engineered to do something large reactors historically struggled with: adjust their output. European Utility Requirements already specify that SMRs should be capable of continuous operation between 50% and 100% of nominal power, with load cycle changes up to 200 times per year. NuScale built a dedicated system for this called NuFollow™, which allows the plant to ramp down during solar-heavy periods and ramp back up for the evening peak. ⚡
Research published in peer-reviewed journals has found that SMR load-following can reduce solar intermittency impacts by 15% to 50% depending on how the systems are configured and the depth of the load-following region the reactor operates in. That’s not trivial. That’s the kind of grid stabilization that currently gets handled by natural gas peakers — the assets that show up at dusk, burn expensive fuel, and produce emissions while they’re at it.
But there’s a more interesting option than just ramping down. When solar floods the midday grid and prices collapse, an SMR doesn’t have to throttle back its reactor core at all. Instead, it can redirect its thermal output:
Hydrogen production via electrolysis, storing clean energy in molecular form
Desalination, producing fresh water when electricity is cheap and demand is low
District heating, banking thermal energy in large heat stores for evening or winter use
Industrial process heat, decarbonizing sectors that can’t easily electrify
The Information Technology and Innovation Foundation’s 2025 analysis makes an elegant point here: a desalination facility coupled to an SMR is, in effect, an economic battery — storing excess energy as clean water when grid prices are low. The same logic applies to hydrogen. An SMR can eat what the solar grid can’t sell, transform it into something useful, and sidestep the operational inefficiency of throttling a nuclear core repeatedly. 💡
Do you think the energy sector is treating nuclear and solar as competitors when they should be engineering them to cooperate? It’s worth thinking about what that shift in framing would mean for investment decisions, grid policy, and the pace of decarbonization.
The economics of coexistence
Price is where this debate gets sharpest. Solar is cheap. SMRs are not — at least not yet. Wood Mackenzie’s 2025 LCOE analysis puts single-axis tracker solar at as low as $37/MWh in the Middle East and Africa. The Darlington SMR project in Ontario, Canada — the first new-generation SMR construction to receive final government approval — comes in at roughly $15 billion USD for 1.2 gigawatts of capacity. That’s more expensive per kilowatt than building a solar farm in the same region.
That comparison, though, measures different things. It’s a bit like comparing the cost of a generator versus the cost of a window: both relate to light and power, but they serve different jobs at different times. A solar panel in Ontario produces electricity meaningfully about 900 hours a year at full output. A nuclear reactor runs at over 90% capacity year-round, including winter nights, cloudy weeks, and the extended shoulder seasons when solar yields fall. 📈
The comparison that matters is the cost of reliable, dispatchable, 24/7 clean power. There, the math shifts. According to TotalEnergies’ analysis, solar-plus-storage at grid scale runs around $100/MWh when you factor in the storage needed to actually firm up the supply. The IEA estimates advanced nuclear at $63/MWh. Neither number is final — SMR costs could fall significantly with scale, while battery costs are dropping fast too. But the directional gap is real, and it matters for planners trying to keep lights on during cold dark winters.
The key facts worth holding in mind:
Solar + 4-hour battery storage costs are forecast to fall below $100/MWh in Europe by 2026 per Wood Mackenzie
Long-duration storage — the kind needed to bridge cloudy weeks — is still far more expensive than short-duration systems
SMR designs are aiming for LCOEs between $50 and $80/MWh, per KPMG and Rolls-Royce SMR’s own targets
An IEA mid-year update found nuclear generation grew 11% in the first half of 2025, alongside solar’s spectacular run, not instead of it
The growth of solar and the growth of nuclear aren’t a zero-sum contest. The IEA recorded them rising simultaneously. That’s the scenario that serious climate analysts think is most likely — not one or the other, but both.
The demand problem nobody wants to talk about
All of this is happening against a backdrop that the solar-vs-nuclear debate usually ignores: electricity demand is growing faster than at any point in decades. 🌍
The Ember Global Electricity Mid-Year Insights for 2025 notes that solar alone met 83% of all new demand growth in the first half of 2025 globally. Remarkable. But here is the catch: demand growth is accelerating because of data centers, electric vehicles, heat pumps, and industrial electrification. Some projections suggest that data center loads alone could add tens of gigawatts of continuous, round-the-clock demand to national grids in the coming years.
Data centers don’t close at night. They don’t take a break during cloudy stretches. They need firm, reliable power, and they’ll pay for it. The FERC data shows that solar capacity factors run around 25%. Nuclear runs above 90%. For an AI hyperscaler signing a 20-year power contract, that difference is not aesthetic — it determines whether the servers stay on.
That’s why companies like Google and Microsoft have been signing nuclear power agreements alongside their solar deals, not instead of them. What they’re building is a portfolio — solar to supply cheap daytime power, nuclear to guarantee the rest. The logic is exactly the same for national grids.
None of this means SMRs are without real problems. Construction costs are high and have frequently surprised on the upside. Regulatory timelines are long. The first Western commercial SMRs are still several years from operation. Critics are right that the promised “cheaper, faster” nuclear moment hasn’t arrived yet. Those concerns deserve honest engagement — not dismissal.
But “SMRs have problems” is not the same argument as “SMRs and solar are enemies.” One is a reasonable caution about an emerging technology; the other is a category error. The question isn’t which clean energy source wins. It’s whether grids built mostly on solar can stay reliable around the clock, in all seasons, without burning natural gas when the sun goes down.
Right now, the honest answer is: not easily, and not cheaply. SMRs — flexible, dispatchable, and able to redirect their thermal output in creative ways — are one of the more interesting candidates to fill that gap. Not the only candidate. But a serious one.
What would it take for your region’s grid operator to start treating SMRs and solar as a combined strategy rather than competing options in a budget fight? That’s probably the right question to start with.