OpenAI Killed Sora. Here Is Why AI Video Is Failing.
OpenAI just killed Sora. The AI video app that was supposed to change everything. The one that hit #1 on the App Store. The one that got Disney to put $1 billion on the table.
Gone. In six months flat.
And honestly? This was not surprising at all. Let me break down what happened, why AI video is struggling as a business, and what this means for the future of AI tools.

What Exactly Happened With Sora?
On March 25, 2026, OpenAI announced they are shutting down Sora. Not scaling it down. Not pausing it. Completely killing it.
Sora launched as a standalone app in September 2025. It was like a TikTok for AI videos. You type a prompt, Sora generates a video. It went viral instantly. Topped the App Store charts. Everyone was talking about it.
Then the numbers started falling. Hard.
The Numbers That Tell The Real Story
Let’s look at what actually happened after the hype died down.
Sora hit 3.3 million downloads in November 2025. That was the peak. By December, downloads dropped 32%. By January 2026, they had crashed another 45%. By February, total downloads were barely 1.1 million. That is a 75% collapse from the peak in just three months.
But downloads are just part of the story. The real problem was the money.
Sora’s entire lifetime revenue from in-app purchases was $2.1 million. That’s it. The total amount of money the app has ever made.
Now compare that to what it costs to run. OpenAI was spending an estimated $15 million per day on compute costs to keep Sora alive. That annualizes to $5.4 billion. Each 10-second video clip costs OpenAI roughly $1.30 to generate and requires 40 minutes of GPU time.
Bill Peebles, the head of the Sora team, put it bluntly. He said the economics are completely unsustainable right now.
Disney’s $1 Billion Deal Is Dead Too
In December 2025, Disney signed a massive deal with OpenAI. They planned to invest $1 billion. The deal would have let Sora users create videos using over 200 Disney, Marvel, Pixar, and Star Wars characters.
Three months later, the deal is dead. No money changed hands. Disney put out a statement saying they respect OpenAI’s decision to exit the video generation business.
This is massive. When a company like Disney walks away from a billion-dollar AI deal, it tells you something about where this technology stands commercially.
Why People Are Not Willing To Pay For AI Video
This is the bigger question everyone should be asking. Sora is not the only AI video tool struggling. The entire space has a payment problem.
Here are the facts.
If you want to use the top AI video tools individually, you are looking at $300+ per month. Sora, Runway, Kling, Pika. Each one wants a separate subscription. That is not affordable for freelancers, creators, or small businesses.
50% of smartphone users in the US said they are not willing to pay extra for AI features. And these are the early adopters. The general population cares even less.
Pika, one of the biggest AI video startups with a $470 million valuation, has a 1.6 star rating on Trustpilot. 87% of reviews are one star. Users complain about confusing credit systems and inconsistent output quality.
The reality is simple. People love playing with AI video for fun. Making memes. Creating weird clips for social media. But when it comes to actually paying for it as a professional tool? The quality is not there yet. The consistency is not there. And the pricing makes no sense for what you get.
AI Video Has A Fundamental Business Problem
Let me explain why this matters beyond just Sora.
AI video generation is the most compute-intensive AI task right now. Text generation is cheap. Image generation is manageable. But video? Every second of video requires processing hundreds of frames. The GPU requirements are insane.
OpenAI’s head of Sora literally said last November that their GPUs were melting. That’s not a metaphor. The infrastructure simply cannot handle AI video at consumer scale without burning through billions of dollars.
And here is the catch. The people generating AI videos are mostly casual users making fun content. They are not enterprise customers paying premium prices. So you have the highest compute costs in AI being served to the lowest-paying customer segment.
That math will never work.
What This Means For The Future Of AI Video
Sora shutting down does not mean AI video is dead forever. But it does mean the current approach is not sustainable.
Here is what I think will happen next.
AI video will move away from standalone apps and become a feature inside existing platforms. Think Adobe adding it to Premiere Pro. Canva is integrating it into its video editor. Not a separate app you pay $20/month for.
The compute costs need to drop dramatically before AI video can be a real business. We are probably 2 to 3 years away from that.
Enterprise use cases will survive. Movie studios, ad agencies, and production houses will keep using AI video tools because they can justify the cost. But the consumer market is not ready.
Open source models like those from Stability AI will fill the gap. If you have your own GPU, you can run video generation locally without paying subscription fees.
Key Takeaways
1. OpenAI shut down Sora on March 24, 2026, after just six months. Downloads crashed 75% from peak.
2. Sora earned only $2.1 million total but cost $15 million per day to run. The economics were impossible.
3. Disney cancelled its $1 billion investment deal after OpenAI pulled the plug.
4. The entire AI video market has a payment problem. People use these tools for fun but refuse to pay premium prices for inconsistent quality.
5. AI video generation is the most compute-heavy AI task. Until GPU costs drop significantly, standalone AI video apps will keep failing.
6. The future of AI video is likely inside existing tools like Adobe and Canva, not standalone apps.
My Take On This
I have been saying this for months. AI tools that burn cash without a clear path to revenue will not survive. Sora is the biggest proof of that.
The hype cycle for AI video was real. Everyone was excited. But excitement does not pay server bills. And when you are spending $15 million a day, you need customers who are willing to pay. Sora never found them.
If you are building a career around AI, focus on tools that solve real problems. Not the flashy demos. Not the viral clips. The tools that businesses actually pay for. That is where the money and the jobs are.
What do you think about this? Drop your thoughts in the comments.



