Why OpenAI’s Sora Crashed and Burned After Just 1.5 Years: The Video Slop That Ate Itself

It was supposed to be the future of creativity. In February 2024, OpenAI unveiled Sora — its text-to-video model that could turn a simple prompt into jaw-dropping, cinematic clips. Hollywood trembled. Creators cheered. Investors poured in.

Fifteen months later, it’s dead.

On March 24, 2026, OpenAI quietly announced it is shutting down the Sora consumer app on April 26 and killing the API by September. The product that once promised to democratize video generation has been euthanized — not because the technology failed, but because the economics were catastrophic.

The Real Killer: It Was Never Profitable

Here’s the uncomfortable truth nobody wanted to say out loud while Sora was still trending: video slop is insanely expensive to run, and almost nobody was willing to pay enough to cover the bill.

Daily burn: roughly $1 million per day (some reports put the total losses in the billions).
Peak users: around one million worldwide.
Active users at the end: fewer than 500,000.

Even at premium pricing, the math simply didn’t work. Generating one minute of high-quality video still costs OpenAI a small fortune in compute. Most users generated a few fun clips, posted them on social media, and never came back. The “wow” factor was real. The repeat usage and willingness to pay… not so much.

Compute Hunger That Starved the Rest of the Company

Sora wasn’t just burning cash — it was devouring the very GPUs needed to train OpenAI’s next-generation models.

Internally, the team was already preparing Sora 3 — a version that executives promised would “blow everyone away” with its creativity and consistency. Sam Altman reportedly ran the numbers himself… and pulled the plug on the entire product line.

The message was clear: keeping a money-losing consumer video toy alive was no longer acceptable when the company needed every watt of compute for higher-priority work.

The Disney Deal That Died in the Spotlight

The timing made it even more brutal.

Just three months earlier, OpenAI had signed a major multiyear partnership with Disney — a $1 billion investment plus licensing deals to bring Disney characters into Sora. The dream of AI-generated Disney princesses, Pixar-style shorts, and Marvel trailers generated by text prompt was inches away from reality.

Disney executives reportedly learned about the shutdown less than an hour before the public announcement. The partnership has now been canceled. The mouse is once again hunting for a new AI video partner — and the broader industry is watching closely.

What’s Next: Everything Moves to “Spud”

The compute that Sora once hogged is being redirected. OpenAI is pouring those resources into a new internal project codenamed Spud — a next-generation frontier model expected to power agents, robotics, and serious business tools rather than viral TikTok clips.

In other words, the slop ate itself. The flashy consumer experiment is gone so the company can double down on the infrastructure and models that actually move the needle on revenue and real-world utility.

Alo read:

The Irony

The craziest part? Sora’s underlying technology wasn’t bad. The videos were often stunning. The problem was never quality — it was unit economics in a world where generating video is still ridiculously compute-heavy.

And while Sora is dead, the hunger for AI-generated video isn’t. Other players (Runway, Kling, Luma, Pika, and a dozen Chinese labs) are still in the game. Disney will simply take its characters elsewhere. The AI princess factory is delayed, not canceled.

Sora lasted 1.5 years. It burned hundreds of millions, consumed precious compute, and ultimately proved that “mind-blowing demo” and “sustainable business” are two very different things.

The slop era of consumer AI video is officially over. The next chapter — more focused, more expensive, and far more pragmatic — is already beginning.