Better Offline AI Is Worse Than The Dot Com Bubble: Part One
46 snips
Jan 27, 2026 A sharp look at why the AI boom may inflict far heavier damage than the dot‑com shakeout. The conversation contrasts past winners and spectacular startup flops. It digs into enormous GPU and data‑center spending, depreciation burdens, and how layoffs hide AI economics. The human costs of market collapses and long recovery timelines are also highlighted.
AI Snips
Chapters
Transcript
Episode notes
Big Tech's GPU Binge Changes The Risk
- The AI bubble differs because major, profitable tech firms are buying massive GPU fleets that create long-term depreciation burdens.
- This hardware-centric spending makes the current bubble far riskier than the dot-com era's mostly speculative startups.
Depreciation Masks True Cash Position
- Depreciation from giant hardware purchases reduces reported net income for years despite being non-cash charges.
- Companies will keep reporting depressed profits while sitting on already-paid but rapidly obsolete GPUs.
Concrete Examples: Microsoft And Amazon
- Microsoft increased PP&E from about $88B to $230B in roughly two years, dragging earnings via depreciation.
- Amazon also buys billions in GPUs and custom chips, which creates similar profit pressure and leads to job cuts.