Frontier AI is getting rejected by businesses | Dean Baker
Summary
Dean Baker argues that the AI industry faces a massive revenue crisis as businesses prefer cheaper, older models over expensive frontier AI. He warns that hyperscalers are overbuilding data centers without clear returns, risking financial collapse. Baker also highlights the superior reliability and privacy of open-source models, which threaten the business models of OpenAI and Anthropic.
Topics discussed
Calshi Prediction Market Ad
Perk Travel and Spend Ad
Intro: AI Industry Revenue Concerns
Salesforce Dreamforce: Slowing AI Appetite
The Revenue Gap: Frontier Models vs. Reality
Hyperscaler Data Center Spending and ROI
Credit Default Swaps and Bankruptcy Risks
Token Price Cuts and Subsidized Usage
AI Safety Risks: Real-World Incidents
Threat of Open-Source Models to Big Tech
Enterprise Preference for Older, Cheaper Models
Sponsor Break: Calshi, Perk, and Num
Customer Retention and the Open Model Shift
Data Center Buildout Sustainability and Tesla Analogy
Sponsor Break: Calshi, Perk, and Num
Pension Fund Incentives and Bubble Dynamics
Potential Triggers for a Market Collapse
Accountability for Money Managers in AI
Timing the Bubble Burst and Lease Commitments
Listen ad-free on Castria