I think the best signal that this massive AI capital investment boom is not a bubble is this: AI agents on the org chart you don’t want to fire.
This is already happening in software development, but I don’t think that’s enough to cover $1 trillion in chip and data center investment.
And this will require more than most business users paying for $200 a month of Claude and Open AI tokens to improve our productivity.
We need to pay real money for real “jobs to be done” in a few years, that:
– don’t go away and continually improve
– humans won’t do or can’t do
– businesses happily pay for in token spend or AI app credit spend
– create direct value in the business
– that don’t get cut when the business gets tight
That might be a tall order, but it’s not impossible.
And I don’t think this is even a net loss for total US employment. So far, there have been enough new jobs added to make up for the AI “jobs lost.”
I think most of this “AI agent spend on our org chart that doesn’t go away” will do important jobs we couldn’t imagine paying someone part-time or full-time to do.
I lived through the dot-com boom and bust in 1999-2001, which cratered hard.
But there were no revenues and profits driving the new Internet businesses that were supposed to pay off the big capital expenditure investment back then.
Now we have useful use, for the most part, and revenues and profits happening at impressive scale. So far.
Most of the massive, unprecedented AI infrastructure buildout is financed with debt.
If the revenue and profits don’t show up to pay this back in the next 3-5 years, it’s going to crater hard.
How do you see this? Bubble or boom?
