I'd go even a step further: For the big corp, having a point of failure that lives outside its structure can be a feature, and not a bug.

"Oh there goes Super Entrepise DB Partner again" turns into a product next fiscal year, that shutdowns the following year because the scope was too big, but at least they tried to make things better.

RTO/RPO is a thing. Despite many companies declare waht they need SLA of five nines and RPO in minutes... this situations are quite evident what many of them are fine with SLA of 95% SLA and PTO of weeks