OpenAI reports $13 billion revenue yet massive losses as it charts personal AGI roadmap

OpenAI disclosed that its 2025 revenue grew to $13.07 billion, up from $3.7 billion in 2024, but operating expenses surged, leading to an adjusted loss of $20.92 billion for the year. Research and development costs rose from $7.81 billion to $19.18 billion, with $10.59 billion paid to Microsoft for cloud infrastructure. Inference and scaling expenses also climbed, pushing the cost of revenue to $7.5 billion. After accounting for a one‑time accounting adjustment, the underlying operating loss is closer to $8 billion. The company said these figures highlight the escalating cost of building and running large‑scale AI models.

In a separate announcement, OpenAI said its next stage is the development of a “personal AGI” – a general‑purpose artificial intelligence system it aims to make available to billions of users. The envisioned system would assist with learning, work, planning, content creation, programming and scientific research. OpenAI did not give a timeline, pricing or geographic rollout details, noting that key decisions on access, safety and control will remain with the company. The dual messages underscore both the financial strain of AI development and the firm’s ambition to expand AI capabilities to a mass market.