OpenAI: Revenue Growth Amidst Massive Development Costs

OpenAI, a leading company in Large Language Model development, is preparing for an initial public offering (IPO), as evidenced by its SEC filings. In this context, recently leaked financial documents offer an in-depth look into its economic situation. These audited financial statements, obtained by independent journalist Ed Zitron and also reviewed by the Financial Times, reveal a company with rapidly expanding revenues that is nonetheless facing even larger expenses.

The data indicates significant revenue growth, increasing from $3.7 billion in 2024 to an impressive $13.07 billion in 2025. By the end of 2025, the company's monthly revenues had already reached nearly $2 billion, suggesting a consistent growth trajectory throughout the year. Despite this expansion, OpenAI's financial sustainability is challenged by the enormous resources dedicated to research and development.

The Impact of Research and Development on Financials

Research and development (R&D) expenses represent OpenAI's most significant cost item, far exceeding total revenues in the last two years. In 2024, R&D expenses amounted to $7.81 billion, already a considerable figure. However, in 2025, this cost surged, reaching an astonishing $19.18 billion.

These numbers reflect the massive investments OpenAI has made in training its new models. A substantial portion of these expenses, totaling $10.59 billion in 2025 alone, was paid to Microsoft, a key partner in development and infrastructure. Training cutting-edge Large Language Models requires immense computational resources, including thousands of high-performance GPUs, dedicated VRAM, and specialized network infrastructures, which translate into extremely high operational and capital costs.

Implications for the Industry and LLM TCO

OpenAI's financial data offers a clear perspective on the scale of investment required to stay at the forefront of LLM development. These extremely high costs have direct implications for the entire industry, influencing deployment strategies and the Total Cost of Ownership (TCO) for companies intending to integrate these technologies. Whether opting for cloud-based solutions or an on-premise deployment, the resources required for training and inference of complex models are a critical factor.

For companies evaluating an on-premise deployment of Large Language Models, these figures underscore the magnitude of investment needed for development and training. The choice between self-hosted infrastructure and cloud services involves a careful analysis of trade-offs, considering factors such as data sovereignty, regulatory compliance, and, naturally, the overall TCO. AI-RADAR offers analytical frameworks on /llm-onpremise to support the evaluation of these alternatives, providing tools to compare the constraints and opportunities of each approach.

Future Outlook and Financial Challenges

OpenAI's financial trajectory highlights a common challenge for companies operating in the generative artificial intelligence sector: balancing accelerated innovation with economic sustainability. Despite rapid revenue growth, the ability to generate net profits is still hampered by colossal R&D investments, which are essential for maintaining a competitive edge.

As the company approaches its stock market debut, managing these development costs will remain a crucial element. The LLM market is constantly evolving, and the need to train increasingly larger and more sophisticated models will continue to demand significant resources, raising questions about the scalability of current and future business models in the AI landscape.