Say promises about AGI. I would expect company to show extensive theoretical work with error bars and such exactly when they can deliver it. And if sufficiently proven for that work to be incorrect or over promising in slightest to fully pay back investors for any losses.
Saying you're hoping to develop some tech that hasn't been done before isn't actually fraud. Everyone knows it's uncertain as to timing and if they'll be able to do it.
