small module nuclear reactors.

Hmm let's see what changed since I last looked. This study seems recent, just looking at the publicly available sections:

SMRs do not represent dramatic improvements in economics compared to large reactors.

Translation: They're way more expensive than renewables. SMRs have some advantage which are mentioned (less land usage, non-intermittency), then we have

The advanced SMRs are compared to conventional large reactors and natural gas plants,

...but not renewables+storage, which would be a good comparison point. If it looked any good they definitely would've included it.

Now that doesn't mean that these things don't make sense for Microsoft. It might e.g. simplify power distribution within datacentres to a degree that other sources just can't, also reduce or eliminate the need for backup power, etc. But generally speaking I'm still smelling techbro BS.