Scalable data models for cross verticals

I was wondering something. How do giant clearinghouses like BlackRock manage multiple verticals, in terms of data modeling, reporting and analysis? Or GE/NBC. A company that has major investments in oil and energy, but also significant investments in retail and logistics, but also heavily invested in media and entertainment. How do they handle the massive cosmos of dimensions and facts, and spin it up into something useful? And for that matter, what does reporting look like? I'm sure it's not just a bunch of different data marts for each niche. Why would you silo everything like that? After all, the whole point of diversifying was to marry up unlikely pairings of your various subsidiaries for innovative new revenue streams. So what does it look like under the hood?