This article examines the financial viability of OpenAI’s GPT-5 model, revealing that while it generates substantial revenue, its overall profitability is questionable. Gross profits may appear healthy, but when operational and R&D costs are factored in, OpenAI likely incurs losses. The findings suggest that AI models may struggle to recoup their development costs within their short lifespan.

The article analyzes the unit economics of large language models (LLMs), focusing on the compute costs associated with training and inference. It discusses how companies like OpenAI and Anthropic manage their financial projections and cash flow, emphasizing the need for revenue growth or reduced training costs to achieve profitability.