No my point is why pay huge sums when if you have all the songs in the world at hand, watermarking and searching those watermarks is trivial. I looked into this a while back and the algorithms/libs to do this extremely accurately are available.
Obviously the problem is getting access to ever song and keeping it up to date with new ones.
It's naive to think that a company that has been operating for decades only has capability that is implementable in a weekend. Surely there are plenty of edge cases with certain songs/genres and how they mix with various environmental noises. That in addition to already having the system scaled, and having a team of people who are intimately familiar with maintaining that system.
Then there's also the userbase. Even if you successfully clone the app, most Shazam users won't even know it exists.
Seems for them the cost (financially and in terms of overhead) of building that team vs. buying a very experienced one was not worth it. They almost have ~$300 billion just sitting around.
The basics might not be that hard, but you can be pretty sure that Shazam has perfected the tech in 15 years of operation while aquiring hundreds of millions of users in the process. Apple also has more money than anyone, it might just make more sense to just buy it and integrate the brand and userbase.
It could just be that owning the goose makes more sense than buying eggs at their volume.