They will need a lot of discipline to not screw up with that much money. It is very hard to be focused, subject your company to real market forces, and to spend money appropriately when it's basically infinite. Where I've seen this happen, a huge management layer grows and exists based on who can do politics better in order to control a chunk of the money, not who is actually moving towards product success. We'll see how these guys do, personally I think it's set up for failure, purely based on the amount of $
Unless most of it's going to train some kind of giant ML model.
That would be super cool if the whole thing was premised on just needing 500 Million in compute.
Anthropic CEO here. It is in fact our intent to primarily spend the funds on compute.
So the person who decides what chips to buy is now everyone's friend.