Because few are motivated enough to tackle 'boring' problems. Who wants to design a real estate database when you could be working on self driving-crypto-cloud-next level bullshit?
Lots of people take on "boring" markets. I don't think that's the problem.
Zillow and Redfin are a few steps removed from being MLS themselves. But for some reason they cannot morph into it.
If MLS threatens to shut them off, that might be grounds for an antitrust lawsuit.
Maybe they're too afraid to risk the business on it? Maybe they did the risk analysis and decided it's not worth it?
Zillow makes money by selling placement on the listings to other agents. Many Realtors really dislike Zillow because of this ("This is MY listing and they are putting someone else's name on it, taking 3% away from me.").
FWIW, I spent time at a large national "disruptor" broker and this is a non-trivial process. There are hundreds of MLS systems in the US, each with its own data, quirks, and synchronization issues. One has to manage so many feeds, cleanup, normalization, etc. Throw in the many data sharing agreements between various MLS systems and the simple act of deduping suddenly gets complicated.
For consumers, the value is that they can see an aggregate list in one place without having to sign up for them via an individual broker's site.