(Photo Illustration by Jaque Silva/NurPhoto via Getty Images)
Traffic to OpenAI’s ChatGPT has been so high that the company is warning it might need to delay future products to accommodate user demand.
OpenAI is experiencing a surge in traffic after ChatGPT added a native image generator that is churning out Studio Ghibli-inspired creations.
"We are getting things under control, but you should expect new releases from OpenAI to be delayed, stuff to break, and for service to sometimes be slow as we deal with capacity challenges," CEO Sam Altman tweeted.
Altman put out a tongue-in-cheek call for more computing power. "Working as fast we can to really get stuff humming; if anyone has GPU capacity in 100k chunks we can get asap please call!" However, the company was confident enough to relax restrictions on ChatGPT's image generation last night, opening it back up to free users.
Still, Altman's tweets suggest that GPT-5, the company’s next-generation AI model, may take longer to arrive than expected. In February, he teased a launch within months. But last month, when OpenAI debuted GPT-4.5, the company limited its availability to paid users after running out of GPUs to run the model.
The GPU scarcity problem raises questions over whether this will be a long-term problem for the company, which is now seeing 500 million people use ChatGPT every week. On the plus side, OpenAI on Monday announced it had raised another $40 billion, but the San Francisco lab has burned through billions to buy and run its GPUs, which require enormous amounts of electricity.
In response, OpenAI's partner, Nvidia, is preparing new GPU architectures that promise to offer performance boosts while improving efficiency. But again, companies will need to spend billions to build out the necessary data centers.