Last Week in AI #333 - ChatGPT Ads, Zhipu+Huawei, Drama at Thinking Machines
Last Week in AI Last Week in AI
OpenAI is testing labeled banner ads in ChatGPT for free and $8/month Go users, starting in the US soon. Translation: the free ride is ending as OpenAI's cash burn catches up with it.
Based on reporting by Last Week in AI, Last Week in AI — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
OpenAI has quietly confirmed what a lot of people suspected was coming: ads inside ChatGPT. The company will start testing banner-style advertisements for logged-in users on the free tier and the $8-a-month ChatGPT Go plan, with a rollout beginning in the U.S. and expanding to other markets over the coming weeks. These won't be sponsored answers woven into the chat itself, at least not yet. Instead they'll show up as clearly blocked-off sections, separate from the actual conversation with the model.
The timing says a lot. OpenAI has been spending at a pace that makes even well-funded rivals nervous, pouring money into compute, data centers, and the endless arms race to keep GPT models ahead of Google, Anthropic, and a growing list of Chinese labs. Free users have never generated meaningful revenue on their own, and the $8 Go tier is priced low enough that it's more about hooking people into the ecosystem than actually covering costs. Ads give OpenAI a way to squeeze some return out of the hundreds of millions of people who use ChatGPT without paying a cent.
It's also a pretty clear signal about where OpenAI thinks its business model is heading long term. Subscriptions and enterprise API revenue are real, but they're not infinite, and the cost of running frontier models keeps climbing. Advertising is the oldest trick in the internet playbook for turning attention into cash, and ChatGPT has attention in spades. The company is betting it can bolt ads onto a product built for direct, personal-feeling conversation without wrecking the trust that made people want to use it in the first place.
Whether that bet pays off depends a lot on execution. Search engines and social platforms spent two decades normalizing ads before people mostly tuned them out. ChatGPT users are still in the honeymoon phase, and shoving banner ads into a tool people use to think through problems, draft emails, or get medical advice is a different kind of intimacy than scrolling a feed. OpenAI clearly knows this, hence the careful framing as separate, labeled sections rather than ads mixed into responses. But once the door opens a crack, it rarely stays that narrow for long.
My take — AI-written commentary, not fact-checked reporting
This is the moment ChatGPT stops feeling like a lab experiment and starts feeling like every other ad-funded product you've ever used, and I don't think that's an accident, it's the bill for burning cash at OpenAI's scale finally coming due. Banner ads today, something closer to sponsored answers tomorrow, that's how every free-to-ad-supported pivot has gone since the dawn of the internet, and I'd bet real money ChatGPT follows the same script within two years.
Read more about this at: Last Week in AI