The infrastructure underpinning AI's growth is colliding with three hard constraints: money, water, and attention. Anthropic's settlement with authors reveals how liability gets redistributed downward rather than absorbed by builders, with publishers now fighting over scraps while the company moves forward. Travis Kalanick's Atoms pivoting toward robotaxis signals that capital follows proven unit economics, not moonshots, and that founders circle back to businesses they know can scale. Apple's Siri AI, meanwhile, languished in beta long enough to become invisible even to early adopters, a reminder that computational sophistication means nothing without the distribution and daily habit to sustain it. The data center backlash is real but misdirected: Americans oppose facilities near their homes regardless of ownership, yet China serves as a convenient villain for policy makers who'd rather blame foreign actors than manage local opposition to the infrastructure their own AI investments require. Long-term rates approaching five percent pose the actual throttle on expansion, not rhetoric about water usage or foreign competition. When capital becomes expensive, only the businesses with clear paths to revenue survive. Everything else gets cut.
Sloane Duvall