NewsNTech
The fundamental tension in AI infrastructure economics is directional.
Building and operating the systems that train and serve large models is capital-intensive and scales with ambition, while the commercial price of AI services has come under sustained pressure.
OpenAI has projected it expects to burn $280 billion by 2030, tying the figure to deeply negative cash flows driven by infrastructure investment and the pricing environment it faces.
The constraint sits in the spread between those two forces. As OpenAI continues to commit capital to infrastructure, its cost base grows.
Keep reading