It doesn’t cost them anything to re-route resources from the keep the lights on work, and has a potentially massive career impact if they’re the first ones to do something sensible with genAI.
Even if most of the work fails it’s still worth it. It’s to their benefit to hype AI beyond all reason because it justifies the resources they spend on it.