People like us think scientists care about big-money things, but they largely don't care about that stuff as much as they care about prestige in their field. Prominent scientists get huge rewards of power and influence, as well as indirect money from leveraging that influence. When you start to think that way, the incentives for fraud become very "minor" and "petty" compared to what you are thinking of.
Indeed, I believe many academic careers were kicked off in this manner. Where it all goes wrong is when other more diligent researchers fail to reproduce said fraudulent research - this is what brought down famous fraudster Jan Hendrik Schön in the field of plastic-based organic electronics, which involved something like 9 papers in Science and Nature. There are good books and documentaries on that one. This will only be getting worse with AI data generation, as most of those frauds were detected by banal data replication, obvious cuts and pastes, etc.
However, when you add a big financial driver, things really go off the rails. A new pharmaceutical brings investors sniffing for a big payout, and cooking data to make the patentable 'discovery' look better than it is is a strong incentive to commit egregious fraud. Bug-eyed greed makes people do foolish things.