For reasons already outlined, I have doubts about their math. Targeting $250/kg payload costs is ambitious for a rocket that has not yet successfully reached orbit or proven cost-effective.

Even if we do somehow succeed at affordably dumping tons of GPUs into orbit, what do we do about the Kessler Syndrome?