AiHub.
Archive view — Fri 02 Oct 2026 Back to live feed Open history

Archived ranking

240 ranked

Inside our months-long investigation into Kevin O’Leary’s Utah data center debacle

Today I’m talking with Josh Dzieza, a longtime features writer here at The Verge, about Kevin O’Leary’s plans to build a massive data center in Utah. The idea was to build the world’s biggest data center — a 40,000-acre AI campus with nine gigawatts of power, or more than double the average power usage of […]

Chips  theverge.com    signal 76  verified

Signals

All →

No signals in the last 7 days.

Just in

Topic radar

All →

Quick takes

Frontier labs are segmenting model families around cost-efficiency, pairing flagship-level coding capabilities with substantial reductions in API token pricing and inference latency.Infrastructure financing is diverging as specialized cloud operators secure GPU-backed debt facilities while commercial banks become more cautious on speculative data center lending.Frontier AI developers are managing contrasting capitalization strategies, with Anthropic preparing public listing milestones while OpenAI continues securing substantial private funding tranches.Frontier AI developers are diverging on liquidity pathways, with Anthropic preparing a late-year public listing while OpenAI defers public markets to secure additional private capital.Lenders and debt markets are tightening underwriting standards for AI data center expansion, forcing infrastructure operators to offer steeper concessions or secure financing directly against hardware assets.Economic feasibility remains the primary bottleneck for physical automation, as research indicates robotic systems can perform the majority of task categories but remain cost-competitive in only a fraction.Frontier lab releases increasingly prioritize inference economics, pairing specialized coding and enterprise capabilities with sharp API price reductions and faster execution speeds across competing model families.Slowing debt appetite from institutional lenders is pushing AI infrastructure developers toward GPU-collateralized borrowing and wider investor concessions to sustain capital-intensive computing expansions.