Victim of success: OpenAI limits free ChatGPT image generation as GPU – and Studio Ghibli – demand soars

OpenAI‘s image generation capabilities in ChatGPT are facing temporary limitations due to overwhelming popularity, highlighting the growing demand for AI visual tools. The company’s decision to cap free users at just three images per day comes as the company struggles with technical capacity, demonstrating both the technical challenges AI companies face when deploying compute-intensive features and the delicate balance between providing free access and managing computational resources.

The big picture: OpenAI is temporarily restricting free ChatGPT users to three AI-generated images per day due to overwhelming demand that’s overwhelming their infrastructure.

• Sam Altman, CEO of OpenAI, explained the situation on X (formerly Twitter): “It’s super fun seeing people love images in ChatGPT, but our GPUs are melting.”
• This limitation comes amid controversy over ChatGPT’s ability to mimic Studio Ghibli‘s distinctive artistic style, raising questions about the balance between creative mimicry and copyright concerns.

Behind the constraints: The image generation capabilities appear to be significantly more advanced than previous iterations, creating unexpected technical challenges.

• Testing indicates the new image generator outperforms DALL-E 3, which was previously used by ChatGPT and will remain the default once users exceed their daily limit.
• Altman acknowledged additional technical problems: “We are refusing some generations that should be allowed; we are fixing these as fast we can.”

Rollout complications: Not all free-tier users currently have access to the image generation tools due to unexpected popularity.

• Altman noted on March 26 that “Images in ChatGPT are way more popular than we expected (and we had pretty high expectations).”
• The image generation feature is already available to all ChatGPT Plus and Pro subscribers, though users report slow performance even on paid tiers.

Why this matters: The temporary limitation reveals the significant infrastructure challenges that even well-funded AI companies face when deploying compute-intensive features to millions of users.

• GPU resources – the specialized hardware that powers AI image generation – represent both a technical and financial bottleneck for companies offering generative AI services.
• The situation highlights the delicate balance companies must strike between providing free access to cutting-edge AI tools and managing computational resources in a sustainable way.