Sam Altman and OpenAI are getting serious.
They’re dropping sidequests like Sora, and trying to catch up to rival Anthropic, which has a booming business selling tools to coders.
But OpenAI still wants to make money from people who will never give it a dime. It wants to do that by showing them ads.
And this chart from analysts at MoffettNathanson explains why:
MoffettNathanson
It’s a simple argument, but I’ll spell it out here: As of January, OpenAI’s ChatGPT had some 900 million users. But the vast majority of them — 850 million — pay very little or nothing at all to use the services.* So OpenAI wants to turn those zero-to-little revenue-generating users into reliable revenue generators by showing them ads.
That’s it. That’s the post.
But, since you are still here: While OpenAI says its barely hatched ad program is already generating results — the company has said it’s on track to generate $100 million a year in revenue, just two months into its ad launch — it still has a very long way to go.
The company is just beginning to build out the team and tech it will need to run a truly meaningful ad business — it just hired a top Meta exec to run sales — and for quite some time, ChatGPT ads are likely to be something ad buyers experiment with, but don’t rely on. Analyst Michael Nathanson says that while the company has been looking to charge advertisers $60 for every 1,000 impressions, it has been settling for something closer to $15 per 1,000 as it gets up and running.
*ChatGPT Go, priced at $8 a month, is OpenAI’s cheapest paid service, and subscribers to that tier will see ads, along with free users.
