I've always enjoyed difficult problems, and working with talented and competent people. I had the chance to study under David Wiltshire, who is probably the most amazing scientist I've ever met. He was the reason I got into cosmology (despite not having studied cosmology or astronomy at all as an undergraduate), and I was quite happy to study just about anything under his supervision. On top of that, he managed to find the perfect project for me, allowing my computational and data analysis skills to complement his focus on the theory.
There was another benefit due to timing. I started my PhD shortly after the start of the pandemic. Thus, the PhD had the added benefit of being three years of guaranteed work without worrying about whatever would happen to the company or their way of working during a global pandemic.
One of the big reasons for ruling out academia is practical. Being an early-career academic requires a ridiculous amount of relocation, and the reward is very low. Pack your bags for two years' postdoc in Europe, two years in some department in Bolivia, perhaps even a third postdoc elsewhere, only to finally (if you're one of the lucky ones) settle down in a permanent position at the National University of Mongolia in Ulaanbaatar, far from anywhere you know after having uprooted two or three times already. Statistically speaking, I'd be unlikely to make it that far, as remarkably few postdocs ever make it to a permanent position, and end up in industry anyway. I've got a family which I'd rather not have to drag between countries, and besides, I grew up here and quite like it.
The other reason is intrinsic to the field. Cosmology was interesting. I had a lot of fun working the the field for four or five years, but I'm not sure I'd do it for forty. It's also quite a low-impact discipline. If I were a condensed matter physicist, my work (or at least work building on my incremental advancement of the field) might be used in some practical materials science application, for semiconductors or something like that. Even if I were a pure mathematician, some fields that were once of no practical use have turned out to be remarkably useful in cryptography or other applications. Cosmology, on the other hand, has no practical use, particularly on the scales I work on. Even in the event that we might travel to another star, I think it's highly unlikely we will ever travel to another galaxy, and many of the things I study are so far away that we are physically prohibited from ever reaching them.
That's not to say that cosmology should not be studied. Blue-skies research has intrisic merit without a practical application to justify it. I'm satisfied with the work I did and the years I spent, and I'm glad I could advance the boundary of human knowledge in a small and focused way. However, I'd much rather work on something where the practical impact of my work is more clearly visible.