Here’s why OpenAI/Anthropic will not go bankrupt (inference margin and training cost).

TLDR: 1. Inference margin is 55~70% for OpenAI and Anthropic. Training model cost should be single digit billions O ($1B). The secured investments are O ($100B). There are a few reasons to be bearish on AI industry that I will list at the end, but most of current AI bears’ thesis are based on misinformation. The bears argue that OAI/ANTH are massively subsidizing the customers and incinerating money they received purely from blind investors. But have they done their research to come up with this