DALL·E drops its waitlist after 2M daily images

Six months after its beta launch, OpenAI is tearing down the velvet rope. The DALL·E waitlist is gone. Starting today, anyone can sign up and start generating images immediately, no more waiting for an email that may or may not arrive. The numbers explain the confidence: more than 1.5 million users are now actively creating over 2 million images every single day. That’s not a typo. Two million. The user base stretches well beyond AI enthusiasts, pulling in artists, creative directors, authors, and architects who are weaving the tool into actual workflows.

The move is a direct result of an iterative deployment strategy that OpenAI has been quietly refining since the DALL·E 2 research paper dropped in April. They’ve treated the beta less like a product launch and more like a public stress test, using real-world feedback to shape the tool. Artists, in particular, get a nod for steering feature development. Their gripes and wishlists led directly to Outpainting, which lets users extend an image beyond its original frame, and a collections feature designed to speed up creative processes. It’s a rare case of a research lab actually listening, and the features show it.

Safety, predictably, was the bottleneck. OpenAI admits as much, saying that learning from real-world use allowed them to improve safety systems enough to justify wider availability. The filters are now more robust at blocking attempts to generate sexual or violent content, and new detection and response techniques are in place to catch misuse. They’re not claiming perfection — the statement carefully says the filters are “more robust,” not impenetrable. That honesty is refreshing and necessary.

What’s next is the API. OpenAI is testing it with select customers and plans a broader rollout for developers and businesses soon. The subtext: they want DALL·E to be infrastructure, not just a consumer toy. The timeline they’ve shared tells the acceleration story plainly — from a paper on GLIDE last November to a million beta users by June to an open door in September. At this pace, the conversation about AI image generation is about to get a whole lot louder.