Dario Amodei explains that artificial intelligence companies must commit spending to chips and data centers several years before knowing what revenue those resources will generate. According to the Anthropic CEO, too little computing capacity could drive customers to competitors, while an excessive investment could make costs unsustainable and potentially lead to insolvency. Anthropic expected an annualized revenue run rate of $8 billion to $10 billion by the end of 2025, after generating about $1 billion in 2024. Its long-term commitments, primarily related to computing capacity and infrastructure, total $900.9 billion through 2045, with $70.7 billion due by September 30, 2027. To limit this risk, Amodei advocates planning around a conservative scenario rather than the most optimistic outcome.
Source: Read the original article
