- OpenAI plans $600 billion compute spending by 2030.
- Company considering IPO valuing firm near $1 trillion.
- OpenAI projects $13 billion revenue for 2025.
- Nvidia nearing $30 billion investment in OpenAI.
OpenAI is intending to invest approximately $600 billion in computing infrastructure by the end of the decade, a person who knows about the issue confirmed to Reuters on the size of the capital it will need to maintain its artificial intelligence goals as it prepares to raise capital through a potential public offering.
The ChatGPT developer is preparing to offer an IPO which could make the firm worth up to 1 trillion, the source reported. The amount being spent is indicative of the increasingly expensive nature of the training of a more complex AI model and the astronomical costs being borne to deploy those systems at large scale.
The source said that OpenAI generates a higher revenue of $13 billion in 2025, compared to the previous forecast of 10 billion, generated in the same year. The amount it expensed in the year amounted to 8 billion as compared to its internal expectations of 9 billion.
The financial news takes place amid Nvidia (NVDA.O) inching closer to a finalising a 30billion deal with OpenAI as part of a massive funding round that might aim to raise over 100billion dollars. With a valuation of around $830 billion (assuming the round is done at the stated valuation), Sam Altman-led company would be one of the biggest privately capitalized companies in the history of the Internet.
Rising Compute Demands
The fact that the investment in compute will amount to $600 billion sheds light on the rising cost of artificial intelligence development. The large language models are huge power guzzlers, not only during the training period, where models are created and optimized, but also during inference, where users are dealing with deployed systems.
The Information separately disclosed that OpenAI had informed the investors that the cost of inferences had increased four times in the year 2025. This led to the reduced adjusted gross margin of the company to 33 percent in 2024 as compared to 40 percent in 2024.
Also Read: Study Finds Most AI Agents Skip or Lack Safety Disclosure Raising Transparency Concerns
The tremendous spike in the inference spending can be attributed to a quick uptake of the products of OpenAI in both the consumer and enterprise markets. Every query, image generation or API request consumes compute resources, which causes recurrent infrastructure costs post-training of models.
OpenAI plans to achieve a total revenue of over 280 billion dollars by 2030 according to CNBC, with approximately half of that amount coming as consumer-facing products and the other half of that figure as enterprise services. That forecast suggests intensive expansion in subscription aspects, corporate incorporations and developer platforms.
Microsoft (MSFT.O) supports openAI and has offered a 3-year deal to supply cloud infrastructure, and investment capital. OpenAI relies on MS cloud software much of its computing capabilities.
Massive Scale Infrastructure.
Even more broad infrastructure ambitions have been previously described by Chief Executive Sam Altman. Altman has claimed in the past year that OpenAI planned to invest 1.4 trillions of dollars in building 30 gigawatts of capacity behind computers, or the equivalent power of a system based on 25 million households in the United States.
The 600 billion compute spend that the source cites is the amount that is estimated to be spent in operation and infrastructure by 2030 as opposed to ecosystem investment.
The spending roadmap is an indication that the race towards AI is now characterized more by access to both capital and energy than novel algorithm breakthroughs. There is a closer relationship between chipmakers and cloud providers as well as AI developers in the race to get access to high-performance processors and data center capacity on a long-term basis.
That would strengthen the relationship between Nvidia and one of its largest AI clients because its upcoming investment of up to $30 billion would place one of the most valuable semiconductor companies in the world even nearer to its large client.
OpenAI refused did not react as yet on the stated numbers, while Nvidia refused to comment on its previous contacts on their investment.
Should the OpenAI undertake an IPO at the valuation of close to a trillion, it would be one of the largest IPOs in technology history as it signals to investors that generative AI services will continue to grow in the long-term despite the massive infrastructure costs needed to support it.
Recommended FAQs
How much does OpenAI plan to spend on compute by 2030?
OpenAI intends to invest about $600 billion in computing infrastructure by the end of the decade. The spending would cover operational and infrastructure costs needed to train and run advanced AI systems.
Is OpenAI preparing for an IPO?
Sources told Reuters that OpenAI is preparing for a potential public offering. The IPO could value the company at up to $1 trillion if investor demand supports that estimate.
Why are OpenAI's compute costs rising so sharply?
Training and running large language models require massive processing power. Increased user activity has also driven up inference costs, which reportedly quadrupled in 2025.
How much revenue is OpenAI generating?
OpenAI is expected to generate about $13 billion in revenue in 2025, higher than earlier forecasts. The company has projected revenue could exceed $280 billion annually by 2030.
How does Nvidia fit into OpenAI's expansion plans?
Nvidia is reportedly close to a $30 billion investment in OpenAI as part of a broader funding round. The partnership would deepen ties between the AI developer and its key chip supplier.
