There is no cloud capacity crunch, what a bunch of hog wash. What we've mostly got are some of the most inefficient systems on the planet where consumers are also building some of the most inefficient projects on top of it.

There is no real effort being put towards maximizing the amount of resources you can get out of hardware, the game for the entire duration of our industry has been throwing more money to buy more hardware.

Now that there are actual limits happening, there still isn't any major motivation to truly rewrite things to be better or more efficient. What they're trying to do instead is force the US government to bail them out via corporate welfare or to threaten allies into buying their wares.

That's certainly another way of looking at it.

But the ground truth is: Cloud customers want to deploy more resources, and they can't build data centers and fill them with hardware fast enough to meet that demand. Cloud customers are fighting for who gets to the front of the line to deploy more stuff. That part is undeniable.

So why don’t these orgs who are voracious for profits just rewrite things to be more efficient? Probably because rewriting things also uses a scarce resource: organizational attention.

I’m in this boat myself on a micro scale: I run a fleet of GPU servers for my SaaS and I know that I could improve the efficiency of what I’m running on them by 20% or more, but it’s far cheaper and more effective in the short run to just throw more hardware at the problem than for me to spend weeks or months overhauling things. At some point the scales may tip, but right now hardware is still cheaper than my time is.