Renewables can't meet the base energy needs. You can't only run your datacenters and factories when the wind blows or the sun shines. They also have low power density, making them problematic at grid scale.

SMRs fix all the issues of modern nuclear reactors. SMR's are not 'small' in the absolute sense, they're on the scale of traditional power plants, not existing nuclear reactors.

They have a ton of advantages:

- They are inherently safe, no need to worry about meltdowns.

- They produce power comparable to existing power plants. Nuclear plants have huge issues with producing tons of power in a centralized manner, meaning the energy infrastructure needs to be designed around them, and probably you need a centralized infrastrucure for power distribution, which might not jive well with local politics. They also need huge concentrated cooling capacity, which might have negative ecological effects, and present a huge risk should they need to be shut down. The recent issues in France with global warming, where the rivers water level got lower and the water warmer, cutting down on cooling margins dramatically leading to shutdowns comes to mind.

- In contrast SMRs can be slotted into current energy infrastructure. Modern reactor designs can be throttled to match grid needs.

- SMRs are standardized, smaller and don't need to be built on site and can be built relatively quicker and cheaper. This is huge. If a traditional plant costs $20B and takes 20 years to build, the interest on the loans could mean it's never going to be financially viable. If you cound do something that makes quarter the power, but costs $5B and 5 years to build, it's an entirely different value proposition.

China is already building these, and they are the main country of origin for solar panels and equipment. Renewables make a ton of sense, but can't solve every issue.

> You can't only run your datacenters and factories when the wind blows or the sun shines.

This has been false in the real world. Factories that use a lot of energy do work with the power company and shutdown all the time based on demand. Large steel mills have their own power plant onsite, but smaller ones just buy grid energy and they want the cheapest. They arrange their factory maintenance schedules with the power company so that the power plants are maintained as the same time they are shutdown. They shutdown every December so the power company can sell the power they were using to run Christmas lights. They often run overnight shifts only because that is when power is cheapest. Even the large factories with their own power generation have shutdown for a couple weeks to sell power to the grid (this is very rare, but it has happened).

Wind and solar is a little more difficult because it isn't as predictable, but that is different from unpredictable. The power companies already are running models to predict the wind and solar cycles because it is important for many things they do. You can bet those smaller factories are already working with the power company to schedule shutdowns when power is predicted to be expensive - factories have to do regular maintenance anyway so it is just a matter of being ready (spare parts) when asked, and wind/solar is predictable enough for this.

> Renewables can't meet the base energy needs.

That assertion is not something everyone agrees with. And baseload is hardly ever qualified with even a ballpark estimate in GW or GWH of capacity needed. So, it's a fairly hollow and meaningless term.

And the reality is that for every 100GW added to grids world wide, about 80% or more is renewable. Nuclear is only small portion of the remaining capacity. And SMRs are a rounding error on that. Most of the rest is gas based generation.

Besides, data centers are a great example of something that can easily scale up and down its energy consumption based on price signals, user demand, etc. So, it's actually ideal to pair with fluctuating supply and demand from renewables. Using e.g. spot instances makes it easy for data centers to scale down their demand if energy is scarce and expensive. Other things they could do is throttle CPUs/GPUs based on energy pricing or encourage people to time shift non critical jobs to when energy is plentiful.

SMRs won't have fixed anything until there are lots of them. Whether you believe this will happen or not, it won't be happening very soon. Realistically, SMRs will remain a niche solution for decades to come; even if they do work at reasonable cost levels.

> You can't only run your datacenters and factories when the wind blows or the sun shines.

You're going to need more work than a bare assertion to demonstrate this, given that storage exists, and given that gas peaking exists, and given that interconnects exist.

Consider these:

- https://www.offgridai.us/

- https://sci-hub.se/10.1039/c7ee03029k

Throttling a reactor makes no sense when the fuel is dirt cheap, which it is for nuclear. It's not clear given the choice of providing the same amount of power with thousands of SMR's worth a few MW's each or a handful of traditional nuclear plants, that SMR's are inherently the better choice. SMR's make obvious sense as a distributed source in cases where power transmission is itself costly and the density of power use is low, but not obviously otherwise.

>- They are inherently safe, no need to worry about meltdowns.

I don't worry about designed meltdowns, I worry about someone bunker-busting it, crashing a plane into it, detonating a convoy of trucks filled with fertilizer inside it, someone deciding that an old mine close the the ground water is "good enough" to store the waste from it, that war, forest fires, floods or famine will leave the sites unmanned until the storage pools dry out and the waste starts burning, or any number of similar scenarios which become so much more likely the more of these things we have.

But mostly I worry that they are more expensive than anything else even in a best-case scenario.

Why should we invest in more expensive electricity, that also carries a significant risk for immense disasters, when we have solved cheap solar, cheap batteries and synthesized fuels? The path forward seems obvious, even though it's less traveled.

but SMRs have the problem of not actually being proven.

China has a few under construction, but having reactors built is not proof of them being viable, e.g. remember the superphénix.