[otterley]

It depends on how you think about spare capacity on your hardware. They're an opportunity cost. Every idle cycle and bit of unallocated memory could be spent doing something else valuable. Consider the most extreme example of GPUs: CoreWeave was founded by repurposing nearly-useless GPUs their founders had on hand to mine Bitcoin (correction: Ethereum) that was obsoleted by ASICs and other specialized mining hardware (correction: transition to proof of stake).

Besides, they did say they were running on EC2, which charges by the second.

[latchkey]

Your CW analogy is wonky since that isn't how it went down. You know my name (I don't know yours. edit: Michael), so you should know a bit more about my history in the space too, right? I can explain it out, but afraid of either being called names, or just not being worth it to you (or me for that matter).

EC2 has preemptable and reserved pricing. It is possible to build autosizing solutions, this is what Google did with AppEngine and later GCP Functions/Cloud Run. Just like optimizing start times, it is also possible to optimize those idle resources. For me, I'd go with the idle resources as the lower hanging fruit over trying to shave ms off making things available on-demand, since it affects the customer experience first.

[otterley]

You're right to call me out on the facts about CW as the story's a bit different from what I remembered from reading an article about them a few months back (https://www.wired.com/story/coreweave-scrappy-cryptominer-mu...).

That said, I think you get the point about idle capital being waste. EC2 Spot Instances are a great example about how to turn that waste into revenue.