Ed Zitron argues that AI is slowing down and that the hype around generative AI is based on deceit and mythology. He points out that OpenAI and Anthropic's massive compute costs require $2 trillion in annual revenue by 2030, but current demand is negligible.
Ed Zitron, in his newsletter 'AI Is Slowing Down', argues that AI progress is decelerating and that the hype around generative AI is based on deceit and mythology. He claims that the infrastructure buildout and compute commitments demand over $2 trillion in annual revenue by 2030, but current demand outside OpenAI and Anthropic is negligible.
Zitron previously published a three-part series on how the AI bubble might collapse. He estimates OpenAI will spend $50 billion on compute in 2026 and Anthropic $30-50 billion, yet both companies lose billions annually. He asserts that there is barely a few billion dollars of demand outside these two money-losing firms.
If Zitron's analysis is correct, the massive capital expenditures in AI may not be justified, risking a bubble burst that could severely impact companies like NVIDIA and data center investors. However, counterarguments exist that AI technology continues to advance, so the debate remains heated.
Comments focus on the debate surrounding the author Ed Zitron's aggressive tone and style rather than the article's claims. Some criticize that his arguments, though fact-based, lack persuasiveness and use the same means as the exaggerated AI propaganda he opposes, while others defend him. Notable opinions include that he appeared calm and factual in a Bloomberg interview and that his predictions have been wrong in the past.