OpenAI goes commercial — cautiously

OpenAI is acting like a company getting ready for Wall Street, not a lab chasing prestige. Chief Financial Officer Sarah Friar said on April 8 that the company plans to reserve part of a future initial public offering for individual investors, which is unusual in big technology listings that mostly favor institutions. (cnbc.com) Friar said OpenAI already tested retail demand in its latest funding round and saw strong interest from individuals. Reuters reported the company has been laying groundwork for an initial public offering that could come as soon as the second half of 2026 and could value the business at up to $1 trillion. (reuters.com) At the same time, OpenAI is building a business that looks more like Google and Meta than a pure software subscription company. Axios reported on April 9 that OpenAI expects about $2.5 billion in advertising revenue in 2026 and sees that number reaching $100 billion by 2030. (axios.com) That forecast matters because ChatGPT started as a paid tool and application programming interface business, where customers buy subscriptions or pay for computing use. Advertising is a different machine: it rewards scale, daily habits, and products that keep people inside the app long enough to show them something sponsored. (axios.com) OpenAI is also trying to avoid the other side of scale, where more powerful models create bigger risks. This week Anthropic unveiled Claude Mythos Preview, a model aimed at cybersecurity work, and limited access to a small group of partners because it could help attackers find software flaws as well as defenders fix them. (cnbc.com) Axios reported that OpenAI is now planning a staggered rollout for some advanced models for the same reason. The company does not want to hand powerful cyber capabilities to the entire internet on day one, so it is leaning toward controlled releases instead of the old pattern of broad launches. (axios.com) Put those pieces together and the strategy looks less contradictory than it first sounds. OpenAI wants mass-market revenue, retail investors, and eventually public-market credibility, but it also wants to show regulators and customers that it can meter access when a model starts to look like a lockpick instead of a chatbot. (reuters.com) (axios.com) That is the new shape of the artificial intelligence business in 2026. The winners are no longer just the companies with the smartest models; they are the ones that can sell ads, survive scrutiny, and decide which tools are safe enough to ship to millions of people and which ones stay behind a gate. (axios.com) (cnbc.com)