undefined | Better HN

0 pointsjsnell3d ago0 comments

The current bottleneck is silicon. Every chip that is manufactured gets housed and powered. (It makes sense: the cost of compute is dominated by capex, the power costs are irrelevant, so they're ok paying a premium for power).

The space data center hypothesis relies on compute supply growing faster than power supply. (Both are bottlenecked on parts of the supply chain that will take ages to scale.)

Even if you believe that's the case, the point at which orbital data centers start making sense is incredibly sensitive to the exact growth rates.

0 comments

lijok3d ago

The current bottleneck is not silicon. There is plenty of silicon locked up in previous gen GPUs that are no longer efficient enough to run relative to newer models. The bottleneck is the economics of owning the older GPU models - which is why all the GPU neoclouds are gonna go bust unless they can get customers to continue renting old GPUs.

The economics are vastly different when opex is near zero for these things

jsnellOP3d ago

All of that is incorrect.

H100 rental prices are still as high as when the cards were brand new. The prices vastly exceed the power costs.

In a world where power or DC permits are the current bottleneck those H100s would be getting retired in favor of Blackwells. But they aren't. They are instead being locked in for years long contracts.

lijok3d ago

Why exactly would the H100s get retired for Blackwells if specifically power and DC permits were the bottleneck?

j / k navigate · click thread line to collapse

0 comments

lijok3d ago

The economics are vastly different when opex is near zero for these things

jsnellOP3d ago

All of that is incorrect.

H100 rental prices are still as high as when the cards were brand new. The prices vastly exceed the power costs.

lijok3d ago

Why exactly would the H100s get retired for Blackwells if specifically power and DC permits were the bottleneck?

j / k navigate · click thread line to collapse