I don't think that was the actual literal expectation, rather that the cost to the tax payer - and there will be a cost to the tax payer - should be best spent elsewhere.
>Two, if they do what they say they want to, they’ll be building new power generation, transmission infrastructures and data centres. Even if AI is a hype, that’s far from useless capital.
Nothing has proven this to be true yet
>OpenAI is spending their own money on this.
Not a single entity has spent any real money on this. So far, its a PR stunt. The general lack of roll out corresponding with the announcement is telling. When real money is spent than I'll believe they might go through with it all the way.
Whats more likely to happen is that these companies will spend at most a token amount of money, then lobby congress and the executive branch for subsidies in order to proceed more 'earnestly' and since this is a pledge, there's nothing in writing that binds a contractual commitment of these funds and their purpose, so they could just as well pocket what they can to offset the costs, use what infrastructure gets built as a result, but shut down the initiative. Bad press won't matter, if its reported on at all.
This has played out for decades like this. Big announcements, so called private money commitments, then come the asks from the government to offset the costs they supposedly pledged to pay anyway, and eventually if you're lucky 1 widget gets built in some economic development area and the companies pocket what they can manage to bilk before its all shut down.