Rohan Paul (@rohanpaul_ai)T2·68 pts
- Schmidt suggests AI may hit a 'money wall' before a 'power wall,' driven by the exorbitant costs of hyperscale data centers.
- With an estimated cost of $50 billion per gigawatt, a 10 GW cluster requires $500 billion, a scale only a few global entities can finance.
- The US capital market's leverage capacity supports current AI infrastructure, but investors will ultimately demand clear ROI.