This is like, the worst possible case for nuclear energy, because this is a random fluctuation that won't be sustained for more than week. Whereas building a nuclear power plant costs billions of dollars and takes multiple decades to pay for itself.
Indeed - traditional nuclear power plants which cost billions of dollars per reactor unit and are wrapped up in mountains of politics are not the future. I think SpaceX has demonstrated the impact rapid iteration, modularization and standardization can have on billion-dollar+ capex projects.
There are SMR technologies that have already passed important phases of regulatory approval. Proposed nuclear technologies, such as from NuScale, will hypothetically be far more responsive to grid conditions than traditional nuclear power plants. Their current design can immediately reject 100% of thermal energy on demand. Power can be added or removed from the grid in increments as small as 77MW with their solution. This level of granularity is easily on-par with existing peaker plant capabilities.
