It could explain his supposed concern about the number of "bots" on twitter. No point in buying a database so that you can create AI to predict the behavior of bots.
But if he did, he'd be an idiot. Which of course is very likely.
I've been exposed to AI for 40 years. The promises have been legendary. It's gone by a variety of names, including machine learning, and expert systems. IBM has had Watson around for at least 20 years. It is a technology with some useful purposes. But the company I worked for tried it about 20 years ago hoping to replace a bunch of engineers. These really smart people came in and interviewed a whole bunch of us. In the end they achieved nothing. I'm pretty sure I knew why too. They interviewed a wide swath of the existing engineers. The really smart ones, the moderately good ones, and the ones that were on the list of the first to go in a downsizing effort. They probably should have only interviewed the really smart ones.
Twitter is a cesspool of idiots with the occasional reasonable person mixed in. The filtering that database would require in order to obtain any predictive value out of a particular population would be approaching impossible. Could politics and advertising in general be able to make use of that database? Maybe, I'm dubious but I suspect it is possible. I'm just not sure it'd be worth the amount he paid for it.
But I'll admit, his focus on people who are writing the software does sort of support your assertion. He may be looking for people talented enough to make use of the database.