Why are OpenAI’s GPUs melting?

OpenAI recently launched a more advanced and accessible version of the image generator capabilities of its popular generative AI chatbot ChatGPT. And people are loving it — a little too much, in fact.

OpenAI CEO Sam Altman posted on X on Thursday that the company was implementing temporary limitations on image generation requests. This is apparently due to the fact that ChatGPT is receiving so many image generation requests that its “GPUs are melting.”

Altman didn’t elaborate on what these “rate limits” would be, only stating that they “hopefully won’t be long!” He also noted that users of the free tier of ChatGPT should be able to do three image generations per day “soon.”