It’s an entirely different skillset to create technology $x than it is to create a successful ad network. Yahoo is the canonical example. It has been one of the most trafficked websites in the world through most of its history and still wasn’t able to successfully sell ads after the dot com bust.
ChatGPT’s revenue means nothing if reports are to be believed that it loses money on each paying customer on just inference. It’s definitely not enough to support its training costs.
Also, I think I remember estimates that it costs 10x as much to serve a ChatGPT result than it does for Google to serve a search result. Not to mention that Google uses its own hardware including TPUs.
Independent analysis: Inference is very profitable. https://martinalderson.com/posts/are-openai-and-anthropic-re... https://www.snellman.net/blog/archive/2025-06-02-llms-are-ch...