Anthropic CEO Dario Amodei faces dotcom-style bubble risk with trillion-dollar AI bet
Amodei's plan to burn $1T on AI may mirror dotcom crash...
A widely shared Reddit post draws a sharp parallel between Anthropic CEO Dario Amodei's current AI strategy and the dotcom bubble's collapse. The author argues that the dotcom crash didn't happen because the internet was invalid, but because startups without a structural moat (e.g., search engines, enterprise software) collapsed when they tried to extract rents without delivering unique value. Today, Amodei is betting that by building superintelligent AI, everyone will have to pay rent and data to Anthropic. To achieve this, he plans to burn trillions of investor dollars on data centers, executive bonuses, and subsidized subscriptions—a play reminiscent of failed dotcom giants.
The critique highlights three core weaknesses in this thesis. First, LLM capabilities are plateauing as human-generated data runs dry; improvements in math and coding are narrow, not general intelligence. Second, the know-how to build and train LLMs is now widely known and replicable. Third, for most real-world money-making tasks, locally run open-source models fine-tuned with good agent design are already “good enough.” Paying $30 per million tokens for GPT-5.5 or Opus offers zero ROI. If Amodei’s trillion-dollar bet lacks a genuine moat, the outcome may mirror the dotcom bubble—a spectacular crash when money runs out and investors demand results.
- Dario Amodei plans to spend $1T on AI infrastructure, subsidized subs, and bonuses, similar to dotcom-era rent-seeking startups.
- LLMs face a data plateau; improvements in math/coding are narrow and don't equate to general intelligence.
- Local open-source models with good agent design already deliver sufficient ROI for real business use, making expensive API subscriptions uneconomical.
Why It Matters
This critique warns that AI's biggest player may be repeating history's biggest financial mistake—no moat, just hype.