According to CNET, OpenAI just launched ChatGPT shopping research for all mobile and web users ahead of Black Friday. The feature uses a GPT-5 mini model that outperforms GPT-5 Thinking on shopping tasks and researches across high-quality review sites and Reddit. It asks clarifying questions and uses conversation memory to create personalized buyer’s guides. Meanwhile, OpenAI faces serious financial pressure with $20 billion annual revenue but $5 billion in 2024 losses, plus projections showing $74 billion in losses by 2028. The company has 800 million weekly users with only 20 million subscribers, and they’re spending $1.4 trillion on data center infrastructure that requires becoming a $1 trillion business by 2030.
The Google factor
Here’s the thing: OpenAI isn’t just building shopping features because they’re nice to have. They’re in a brutal race against Google’s AI progress, and shopping is one of the few areas where they might actually have an edge. Think about it – nearly 60% of Americans now use AI tools for online shopping, and 1 in 4 say ChatGPT beats Google for product research. That’s huge. But can they actually monetize this advantage?
Does this actually work?
The demo results weren’t exactly mind-blowing. When reporters tested it, ChatGPT recommended an air fryer that wasn’t the best option and suggested only Mickey Mouse-themed gifts for a dad. So we’re not talking about some revolutionary shopping genius here. It’s basically a slightly smarter version of what we’ve seen before. And without affiliate revenue or instant checkout, what’s the business model? OpenAI says they’re not taking cuts from retailers yet, but let’s be real – that’s probably the endgame. They need revenue streams, and fast.
The money problem
Now for the scary part. OpenAI is burning cash at an alarming rate. They need to become a $1 trillion business to justify their infrastructure spending, but they’re losing billions annually. Their subscriber growth is decent – nearly tripling to 15.5 million in 2024 – but is that enough? Probably not. Shopping features might help convert more free users to paying customers, but they need to actually work well first. And based on the demo performance, there’s still work to do.
Where this is headed
So what does this mean for the rest of us? Basically, we’re about to see AI shopping assistants become the new normal. The timing with Black Friday is no accident – they want maximum visibility. But the real test will be whether people actually trust ChatGPT’s recommendations over their own research or established review sites. And for OpenAI, the pressure is immense. They’ve built this incredible technology, but now they need to prove it can actually make money. Otherwise, those trillion-dollar dreams might remain just that – dreams.
