LLMs slide on runaway capex and thin revenue for five straight days. Why? Gravity
Wall Street is clear
LLMs extended a five-day selloff, even as the NASDAQ was propped up by chipmakers.
But flat adoption rates and pre-execute provisioning efficiencies point to falling demand for AI infrastructure.
History is clear
From the 1873 railroad crash to the 2000 dot-com bubble, when capex races ahead of adoption, narrative eventually gives way to arithmetic.
Seeing red clearly
Real-world data from the field
Enterprise adoption remains a challenge according to MIT…
- For Microsoft
- For Salesforce
- For Amazon
The warning signs are no longer theoretical
- Adoption has stalled at all LLMs in December and for all of 2025
- LLMs have failed to report flat adoption rates on 10-Qs
- Salesforce has retrenched, admits AI is not working
- AI-powered Alexa “killed my fish”: Tales from Amazon employee beta testers
- MIT: 95% of all AI-for-business implementations fail
- Menlo Ventures: ChatGPT only converts 5% from free to paid
- Unrealistic valuation: Anthropic’s $350B is based on $11B investment
- OpenAI considers advertising, fails to monetize world’s largest AI user base
- Scale has failed to produce revenue commensurate with capex
- WSJ’s James Mackintosh called the AI bubble
- Jim Cramer: The year of magical investing is over
- Token-based pricing is inaccurate
- Inference costs scale up with adoption, not down
- Rick Perry’s datacenter developer Fermi America lost 40% after losing a tenant
- Oracle is down 40%, losing $25oB in market cap
- Do the math: Flat-rate pricing will never work
- 95% use ChatGPT for free