OpenAI's pivot to commerce: ChatGPT becomes a shopping mall

We don't know yet how OpenAI plans to handle potential conflicts of interest here. If I ask ChatGPT where to buy groceries, will it recommend Instacart because it's genuinely the best option, or because there's a revenue share agreement? The company hasn't disclosed the financial terms of these partnerships, and I couldn't find any public information about how recommendations will be weighted. This is the kind of thing that seems minor now but tends to become a big deal later.

The Reddit partnership is a different beast entirely. Reddit's content, those millions of forum posts and comments and arguments about whether a hot dog is a sandwich, is now training data and source material for ChatGPT. This is valuable because Reddit contains actual human opinions and experiences in a way that most of the web doesn't anymore (thanks largely to SEO content farms and, yes, AI-generated slop). But it also raises questions about what happens when the AI is simultaneously learning from Reddit users and potentially competing with Reddit for their attention.

I keep coming back to the Apple deal because it's the one that makes the most strategic sense and also the one that worries me least. Apple has always been better than most tech companies at keeping user data relatively private, mostly because their business model doesn't depend on selling that data. If ChatGPT is going to be embedded somewhere, an Apple device is probably the least bad option. But it does mean that OpenAI's technology is now deeply integrated into the daily lives of hundreds of millions of iPhone users, which is a lot of influence for a company that's still figuring out its own identity.

Here's what I think is actually happening, and I could be wrong about this, but what do I know. OpenAI burned through enormous amounts of money building GPT-4 and GPT-5, and the computing costs for running these models at scale are genuinely staggering. The company needs revenue, lots of it, and fast. Enterprise subscriptions help but probably aren't enough. The choices are basically: raise more money from investors (which comes with its own strings), charge users more (which limits adoption), or find ways to monetize free users through ads and commerce partnerships.

They've chosen door number three, and I understand why. It's probably the path that lets them maintain the broadest user base while generating the revenue they need to keep the lights on and the GPUs humming. The problem is that this path has a well-documented tendency to corrupt the products it touches. Not immediately, not dramatically, but slowly, in ways that are hard to notice until you look up one day and realize the thing you're using isn't really serving you anymore.

The young founders building AI startups right now probably don't remember when Google's search results were actually good, when you could find what you were looking for without scrolling past four ads and a bunch of SEO-optimized garbage. They don't remember when Facebook was a useful way to keep up with friends instead of an engagement-maximizing outrage machine. They don't remember when these companies made the exact same promises OpenAI is making now about protecting users and maintaining quality.

I'm not saying OpenAI will definitely follow the same trajectory. Maybe they've learned from those examples. Maybe the leadership genuinely cares about keeping the product good even as they monetize it. Maybe the structural incentives are different this time somehow.

But I've been doing this long enough to be skeptical, and the pattern is hard to ignore. First come the partnerships. Then come the ads. Then come the algorithmic tweaks to maximize engagement. Then come the quarterly earnings calls where executives explain why this quarter's numbers didn't meet expectations and what they're doing to fix it. And somewhere along the way, the original mission gets quietly filed away in a drawer somewhere while everyone focuses on the metrics that actually matter to shareholders.

The thing that makes this particularly concerning in the AI context is that ChatGPT isn't just a search engine or a social network. It's increasingly becoming an interface for how people think about and interact with information itself. When your AI assistant has financial relationships with certain retailers and content providers, that shapes what information you receive and how you receive it in ways that are much harder to detect than a clearly labeled ad.

I don't have a tidy conclusion here because the situation is genuinely uncertain. OpenAI might thread the needle and build a sustainable business without compromising the product. The partnerships might remain benign integrations that just make shopping slightly more convenient. The ads might stay clearly separated from the actual AI responses.

Or we might be watching the early stages of another tech company following the well-worn path from idealistic startup to ad-supported platform to whatever comes after that. We've seen this enough times to know how it usually goes.

If you want to argue with me about this, my email's on the about page. I still prefer it to the comments section, which probably tells you something about how optimistic I am about the direction of online discourse generally.

The question isn't whether OpenAI will monetize ChatGPT. They already are, and they probably have to. The question is whether they can do it without losing the thing that made the product valuable in the first place. History suggests that's harder than it looks, but I genuinely hope they prove me wrong. We could use at least one example of a tech company that didn't eventually sell out its users for ad revenue.

I'm just not betting on it.

Sarah Williams · 50 mins ago · 7 min