This is all about web scraping. An AI or even agentic browser, isn’t going to be the same target demo as Chrome users, and browsers are only super lucrative with certain daily drivers; where as AI queries have a much larger cost. So why would OpenAI do this? To train their models, they scrape the web. Cloudflare recently cut them off from doing this at scale with their aggressive bots. So they’ll do it in a browser. Cloudflare treats a user initiated action different from a bot one. So if you type your search and OpenAI goes and finds the answer, not only can it meet Cloudflares crawl rules, but it can save the content it displays to you, and from any site you visit. Turning the end user into a web scraper. This won’t eat into mainstream marketshare and it’s a big loss leader. But it’s critical for them to be able to keep training new models.