OpenAI just put out some jaw-dropping figures that reveal just how much it costs to be at the very pinnacle of the artificial intelligence business. The ChatGPT parent firm now thinks it will cost a staggering $115 billion by 2029 that’s a staggering $80 billion above the original forecast.
Just put it this way: OpenAI’s new budget constraint means the new renovation cost for your house will be three times the original quote, only instead of renovating the kitchen, the new budget goes towards the development of the future of AI.
The ramp-up in spending is quite staggering. OpenAI will burn through more than $8 billion in 2025 alone, already $1.5 billion above their previous estimate. And that’s not the end of the story. By 2026, they’re going to spend upwards of $17 billion close to twice the previous estimate of $10 billion. The figures continue to rise: $35 billion in 2027 and $45 billion in 2028.
So why the massive cash burn? It appears running ultra-high-end AI systems takes a lot of money and continues to do so as the tech advances and refines itself.
The Strategic Investments of OpenAI and Revenue Growth
Most of the budget goes to infrastructure. OpenAI is actually building its own tech empire by investing heavily in the locations for data centers and custom-built server chips. It’s all about saving on third-party cloud providers—buying your own factory instead of paying rent in someone else’s.
Processing costs alone will consume upwards of $150 billion during the period 2025-2030. Why? It takes astronomical amounts of processing power to train and run cutting-edge AI models. As ChatGPT gives the response to a question or produces some text, it uses computational power that doesn’t exactly come on the cheap.
The company also must compete vigorously for the best AI talent and will end up paying some $20 billion in stock compensation by 2030. When everybody wants the best-of-the-best engineers and researchers, you’re going to pay the best prices for them.
OpenAI won’t be satisfied with simply leasing computer power from someone else. It is cooperating with corporations such as Broadcom to build proprietary AI chips and control costs and performance better. It also undertakes giant projects such as the “Stargate,” a prospective $500 billion data center project with as much as 10 gigawatts of capacity.
The firm has extended collaborations with Oracle and Google Cloud and is looking at the prospect of ultimately providing their own cloud services to other firms, like Amazon Web Services became a giant business for Amazon.
Fortunately for OpenAI, the dollars aren’t just going out, the dollars are flowing in faster than anticipated too. The company puts the 2025 total revenue at $13 billion and a staggering $200 billion by 2030, 15% higher than previous estimates.
OpenAI’s Ambitious Revenue Goals and the Trillion-Dollar Price of Staying on Top
ChatGPT subscriptions and commercial partnerships are performing well, with the company expecting nearly $70 billion more revenue over six years compared to earlier projections. They’re also exploring new revenue streams, including potential shopping services and advertising for free users, which could generate $110 billion between 2026 and 2030.
Though $115 billion seems exorbitant, OpenAI is only playing catch-up to the tech titans. It costs Meta alone $70 billion in 2025 to fund AI development projects, Microsoft invests $80 billion, and Amazon spends $100 billion. The entire technology industry can end up spending trillions on AI-related infrastructure in the years to come.
This spree of expenditure shows the intentions of OpenAI to stay competitive in the rapid-changing AI world. But what comes through equally clearly in the process is the unprecedented financial risk in leading the AI revolution. The firm needs to continue to raise money and provide spot-on execution to finance such massive spends.
The bottom line? OpenAI has high hopes for the future of AI and the cost of remaining the leader just continues to get larger.
