The rising demand for AI-generated media has prompted OpenAI and Google to introduce new daily usage restrictions across their image and video creation platforms. The move follows a surge in public interest around Google’s Nano Banana Pro image generator and OpenAI’s Sora video model.
OpenAI confirmed that free users can now generate only six videos per day on the Sora app. Paid users — including ChatGPT Plus and Pro subscribers — remain unaffected, while additional generation credits will be available for purchase. “We’re setting usage limits for free users to 6 gens/day… Our GPUs are melting,” Bill Peebles, OpenAI’s head of Sora, wrote in a post on X. He later clarified that the intent was to make the tool accessible to more people without overwhelming compute resources.
GPUs serve as the primary processing chips powering large-scale AI systems, enabling training and real-time generation. Although hardware components do degrade with heavy use, data-center GPUs are designed to operate under extreme workloads. OpenAI CEO Sam Altman previously described “melting GPUs” during a viral AI boom, later telling Bloomberg the phrase was meant figuratively, not literally.
Meanwhile, Google has also reduced access to its new Nano Banana Pro image generator for free users. Those without a Gemini AI subscription are now restricted to two image generations or edits per day. Google’s support documentation notes that daily limits may change frequently for free-tier access to Gemini 3 Pro Thinking, while the earlier Nano Banana model based on Gemini 2.5 still permits up to 100 images per day.
The scale-back comes during a U.S. holiday weekend, a period historically associated with increased AI tool activity. The restrictions highlight ongoing pressure on energy and compute capacity as generative media platforms expand worldwide.
