I've used a few different AI review tools and I have very mixed feelings about this.

There's an inherent conflict of interest around productized AI that doesn't get discussed enough when you can't fully control the configuration of the AI stack, which models are being used, what the prompts look like, etc. And that causes major pain for a mission critical process like code review.

And this announcement doesn't really make clear how much of this money is going directly to maintainers vs. being offered as free compute to use their product. My guess is that 95+% is going to be given out as credit towards reviews, which is not the same as actually committing money to open source projects through actual donations.