OpenAI spending just hit 4 billion in one year, while the company brought in only
3 billion in revenue. Most of it went into research, marketing, and the giant computers that run ChatGPT. And the company is heading into a public listing valued above one trillion dollars.
The Gist
- OpenAI spending hit $34 billion in a year, mostly research and computing power.
- Revenue reached $13 billion, with monthly income now around $2 billion.
- The company is preparing to go public at a valuation north of one trillion dollars.
Have ChatGPT Recap This Article
ChatGPTWhere the $34 billion actually went
OpenAI is the company behind ChatGPT, the chatbot most people have heard about. Behind that simple chat window sits one of the most expensive operations in tech today.
Over the past year, OpenAI spending reached $34 billion. To put that number in plain words, that is the kind of figure a mid-size country spends on its entire defense budget. And it goes into one company building AI tools.
The single biggest line is R&D at $19 billion. That covers researchers, engineers, and the huge bills for training new models. Training a model like GPT means renting thousands of specialized computers (GPUs, which are chips originally built for video games and now used to run AI) for weeks at a time.
Sales and marketing eats another $6 billion. That is the money OpenAI spends to get businesses to sign multi-year contracts and to keep ChatGPT visible in front of millions of new users. The rest of the $34 billion goes into infrastructure, salaries, and long-term deals with cloud providers to secure compute capacity.
The same pattern showed up earlier this year when AI spending hit $7,500 per employee at top firms. Companies are pouring cash into AI tools, and the tool makers are pouring even more cash into running those tools.
The $39 billion loss and the trick behind it
Here is the part that sounds dramatic. OpenAI’s net loss for the year hit $39 billion, up from $5 billion the year before. On paper, the company is losing money at a record pace.
But a big chunk of that loss is not real money leaving the company. About $30 billion is a non-cash accounting charge. That is an accounting term for a number that shows up on the books but does not actually empty the bank account. It comes from the way OpenAI reorganized its corporate structure earlier.
Take that accounting charge out, and the real cash loss is around $8 billion. Smaller, but still enormous. Eight billion dollars in real money is what most large companies make in profit, not what they lose.
Why does this matter for regular users of ChatGPT? Because companies that lose this much money have to find ways to bring in more revenue eventually. That usually means new paid features, higher subscription tiers, or more aggressive limits on the free version.
The same pressure explains why OpenAI keeps building giant facilities like the Ohio data center with Nvidia. More compute capacity means more users served, but also more money spent. The race to scale is also a race to spend.
Keep learning on AI Noobies:
- Sarvam is India’s newest AI unicorn at $1.5B
- Facebook’s New AI Reads All Your Public Posts
- Google’s New Format Lets AI Agents Share Knowledge
The trillion-dollar IPO ahead
Despite all this, OpenAI is preparing to go public. An IPO (initial public offering, when a private company starts selling shares on the stock market) is in the works. The target valuation sits above one trillion dollars.
A trillion dollars puts OpenAI in the same league as Apple, Microsoft, and Nvidia. It would be the largest IPO ever for an AI company, by a wide margin. And it would happen at a moment when the company posts $8 billion in real cash losses.
Short term, expect more pricing changes in the products you use. Paid ChatGPT plans will keep evolving. Free limits will get adjusted. Business plans will get pushed harder. Every small tweak on the product side comes from this giant financial picture.
Medium term, the IPO will set the tone for the whole AI industry. If investors buy into the trillion-dollar story, every other AI company gets a free pass to spend big. If the market reacts coldly to the OpenAI spending pattern, the financing tap could tighten across the sector.
For now, the only practical takeaway for users is this. The tools you use are running on the most expensive infrastructure ever built. Someone has to pay for it. And in the coming months, OpenAI will be deciding exactly who pays, and how much.
Follow the story on AI Noobies.
About the Author
