I would hand back a grant that was given due to a false result that occurred because of my error. Julia should be praised for doing the right thing with the paper, I just think she should think about going further.
It would be more like losing the remaining payments of a multi-year contract because it gets broken. Or in your analogy, like losing your job.
Sometimes there are going to be mistakes.
But if you sold me software to solve a problem and I paid you for it. You wouldn't refund me if the software didn't solve my original problem? I feel like that's scam but I could be wrong. I didn't get what I was promised and hence I am entitled to a refund.
If you got a laptop with broken keyboard or the privacy friendly app that you paid extra for turned out to be facebook, would you not want your money back?
You can't just use mistake as an excuse if you promise "specific" things like written in your implicit/explicit contract.
Many countries have pro-consumer laws which protect them from exactly this.
That this researcher was honest makes their future research twice as productive in my eyes. Not only are they less likely to make that kind of mistake again, you can probably trust them to not lie about their research.
I think you are correct to want remedy. A big fix, a refund, a workaround, an apology, a million dollars, or a withdrawal of the study. Depending on circumstances.
In my world, most of the implicit contracts are. Shit happens and try to fix it. Yours might be refund me.
Apart from that many great discoveries were made based on initial mistakes.