Let’s be real for a second: OpenAI built ChatGPT, arguably the most recognizable AI product ever. It’s partnered with Microsoft. It’s pushing boundaries with every new model release.
But let’s also be honest: it might never become a $100 billion-a-year revenue juggernaut - and that’s not just some spicy take, it’s grounded in reality.
OpenAI isn’t the only player in town. Google, Meta, Amazon, Anthropic, Cohere, and a dozen open-source communities are all in this race. And some of them (like Google and Amazon) already control the pipes - search, cloud infrastructure, devices, etc.
Plus, open-source models are catching up fast. Why pay OpenAI a premium when you can just run your own AI in-house or on cheaper services? Remember DeepSeek?
Training a GPT-4 or GPT-5-type model costs hundreds of millions of dollars. And that’s just to train it. Every time you ask a question in ChatGPT or call the API, it burns real cash via GPUs and power.
Sure, inference is getting cheaper, but not cheap enough to serve billions of users at scale without insane overhead.
Let’s put it in perspective: only a handful of companies ever hit that kind of annual revenue - Apple, Google, Amazon, Microsoft. That’s because they control gigantic ecosystems (iPhones, Ads, Cloud, Windows, etc.).
AI chatbots and dev tools? Not that big of a market, at least not yet. Most people won’t spend big money for general-purpose AI. They want specific tools for specific jobs - and many of those tools will be cheaper, or even free.
AI is under a microscope. Between copyright lawsuits, misinformation fears, and privacy concerns, governments are getting involved. That means tighter rules, slower innovation, and possibly limited ability to monetize at scale.
If laws end up restricting training data, model outputs, or how models can be used in industries like healthcare or finance, OpenAI’s revenue ceiling gets lower.
Let’s face it: ChatGPT is cool, but it’s not the iPhone. People use it, sure, but is it a must-have tool for everyone? Not really.
Most regular users treat it like a smart toy or occasional helper. The real sticky apps (like Gmail, Photoshop, Excel, or TikTok) solve very specific problems. General-purpose AI doesn’t have that same stickiness - at least not today.
OpenAI’s biggest backer (and customer) is Microsoft. That’s great for stability, but it’s also a double-edged sword. Microsoft is already integrating OpenAI into its own products (like Copilot), which means they’ll likely capture most of the value.
OpenAI becomes more like a backend vendor than a standalone tech titan.
And who knows what Microsoft decides tomorrow? OpenAI can control only so much.
None of these revenue streams alone feels like a path to $100B. And bundling all of them together? That just makes the business complex - and harder to scale cleanly.
OpenAI is good. But that doesn’t guarantee it becomes the next Apple or Google. It could still be a successful company while making less than $100 billion a year provided they figure out to cust costs. And honestly, that might be just fine. However, most likely OpenAI will bite the dust.