The real reason is because they are detailed supply-side models and so that was what we knew best & could do quickly with limited funding

And to expand slightly: Even when I started working with IAMs in 2014, people like Charlie Wilson were involved for a bit already trying to improve demand-side aspects in projects.

I'm also showing my modelling ignorance here but - aren't there quite a prominent suite of 'low demand' scenarios included in the latest round of IPCC reports? The IMP-LD things?

This is why I tend to focus on the systemic characteristics rather than personal recriminations! The impression I get is many of those pressure still very much exist

Yes that's the instant reason, but the set of structures and power relationships that led to supply-side models being more readily available and understood is worthy of interrogation. There's a reason and a history behind why those were the models available that's worth study.