There is another way to look at this...
If capex growth were to slow, the reason will need to be determined.
If hyperscalers stop spending because the demand and returns aren't there, I would agree. And that would be a problem well beyond the companies supplying the buildout...
I’ve written recently about another possibility which is this - monetisation keeps improving while capex growth slows. So the same spending that shows up as revenue for a supplier is cash leaving the hyperscaler. Those businesses don't necessarily need the same outcome from here.
Slower capex growth alone won't necessarily fix FCF especially if the absolute bill were to keep rising. But... if customers keep spending and the infrastructure already built starts earning a better return, the companies paying those bills could become a lot more attractive.
When the hyperscalers stop, in terms of spending cash flow on the AI buildout things will be complicated. But as long as the music is playing, you've got to get up and dance.