>>16771309
> chemical factories that could use the high temperature process heat
There are NPPs for that purpose being built or in advanced stages of planning, and such NPPs are not small. For example, Xuwei NPP whose main purpose will be to supply steam to the nearby Lianyungang petrochemical complex. It will have 4x 1200MWe(3180MWth) pressurized water reactors and 12x 100MWe(250MWth) high-temperature gas-cooled pebble-bed reactors, for a total 6,000MWe(15,720MWth). So not a small NPP. It makes little sense to expend nuclear industry resources on building small NPPs until you have satisfied the need for large NPPs, a need which is widespread. You can use coal or gas in the small cases.

The reason the HTGR reactors are small is due to technical limitations. The Chinese designers originally wanted bigger reactors because it would've been more economical, but they found it would not be technologically feasible because IIRC it would have required the core and the pebble funnel mechanism to have complex geometry, so it was considered easier to just build multiple separate reactors with cylindrical geometry. Diameter is constrained, because control rods need to be around the outside of the core, since they cannot be inserted directly into the core without risking cracking the pebbles, which was a problem with the old German HTGR. Pebble structural strength limits how tall the reactor can be. And passive safety requirements limit the core power density.

>there's the AI slop that need the stable power
They would likely be better off by being connected to a large grid. Normal civilian nuclear reactors need to be regularly shut down for refueling. There are some reactor designs that can handle online refueling, but such designs aren't technologically mature enough that they wouldn't have frequent outages for decades. It wasn't until the 1980s that the US light-water reactors attained a decent capacity factor.