3 Minutes
Six months. That’s all it took for one of the most talked-about AI video apps to rise, peak, and disappear.
When OpenAI launched Sora as a standalone app last September, it didn’t just enter the market—it stormed it. Within days, it climbed to the top of the App Store charts. Within a week, it crossed a million downloads. For a brief moment, it looked like OpenAI had cracked consumer video creation at scale.
Now, it’s over.
In a short post on X, OpenAI confirmed the shutdown. No long explanation, no dramatic send-off—just a quiet goodbye. Behind the scenes, though, the reasoning is less mysterious. Running Sora wasn’t just expensive. It was brutally expensive.
When viral growth meets real-world costs
At its peak, Sora hit around 3.3 million downloads in November. Then came the drop. December saw a 32% decline. January plunged even further, down 45%. The early excitement faded faster than expected.
Revenue told a similar story. Across its entire lifespan, the app generated an estimated $2.1 million in in-app purchases. That might sound respectable—until you stack it against the reported $15 million per day it took to run the system. At that scale, success isn’t measured in downloads. It’s measured in sustainability.
The math simply didn’t work.
And there was more than just financial pressure.
Sora quickly ran into familiar AI pitfalls: deepfakes, copyright violations, and content moderation headaches. Users generated videos featuring real people—including deceased public figures—alongside copyrighted characters like Mario and Pikachu, despite safeguards. The technology impressed, but the guardrails struggled to keep up.
A pivot, not a retreat
The shutdown doesn’t mean OpenAI is walking away from video entirely. Instead, it’s shifting focus. The team behind Sora will now work on “world simulation” research—a move tied closely to robotics and advanced AI environments.
At the same time, OpenAI is simplifying its product lineup. Rather than juggling separate tools, the company is merging its browser, ChatGPT app, and coding features into a more unified desktop experience.
There’s also a quiet casualty in all of this: the Disney partnership. Announced with fanfare just months ago, the deal promised access to over 200 characters from franchises like Marvel and Star Wars, alongside a planned $1 billion investment. In reality, the agreement never fully materialized. No funds were exchanged before the project collapsed.
Disney’s response was measured, almost diplomatic—acknowledging OpenAI’s decision to shift priorities rather than pushing back.
In hindsight, Sora feels less like a failure and more like an experiment that scaled too quickly. It proved demand exists for AI-generated video. It also exposed just how costly—and complicated—it is to deliver that experience responsibly.
Sometimes the most important signal isn’t a product launch. It’s how quickly a company is willing to pull the plug when the numbers don’t add up.
Comments
Marius
Is the $15M/day figure legit? Sounds wild, maybe counted peak infra costs? Either way, running gen-video at scale was always gonna be brutal. If Disney bailed, that says a lot
atomwave
Wow, all that hype for just six months? Crazy how fast the costs killed it. Impressive tech, but moderation and $15M/day... yikes. Curious what they'll build next
Leave a Comment