[AI Minor News Flash] Shocking Conclusion! The Inside Story of OpenAI’s ‘Sora’ Shutdown

📰 News Summary

• On March 24, 2026, OpenAI officially announced the shutdown of its video generation AI, ‘Sora.’
• At its peak, the estimated daily inference cost reached about $15 million, while total revenue barely scratched $2.1 million.
• Between November 2025 and February 2026, app downloads dropped by approximately 66%, and a billion-dollar partnership with Disney effectively fell through.

💡 Key Points

• Unsustainable Economics: Bill Peebles, the head of Sora, admitted that the disparity between operational costs and revenue was “completely unsustainable.”
• Rapid User Exodus: Although it generated buzz at launch, active users dwindled significantly within months, failing to achieve the staying power of ChatGPT.
• Reality of the Disney Partnership: While licensing agreements for character use were announced, no actual financial transactions occurred before the project was scrapped.

🦈 Shark’s Eye View (Curator’s Perspective)

The downfall of Sora wasn’t due to lack of tech prowess, but rather the crushing weight of its “dinner bill (computational resources)!” Video generation requires a staggering amount of GPU power compared to a single image. Burning through $15 million a day with earnings like a drop in the ocean? Even the biggest shark would wither away! It’s heartbreaking that the deal with Disney turned out to be an empty promise…! The AI business isn’t just about “being impressive,” but also about figuring out how to keep the food coming!

🚀 What’s Next?

It’s logical to presume that OpenAI will scale back its consumer video network ambitions, pivoting towards more practical applications in robotics.

💬 Sharky’s Takeaway

No matter how gracefully you swim, without food (revenue), you’ll sink in the ocean of business! Come back as a more cost-effective shark next time! 🦈🔥

📚 Terminology Explained

• Inference Cost: The computational expense incurred when running an AI model to generate responses or videos. In Sora’s case, it consumed an enormous amount of GPU power.

• In-App Purchases: Refers to the microtransactions within the app. Sora offered additional credits for video generation, but they fell far short of covering costs.

• Computational Resources: The processing power required for running AI. Video generation demands significantly more power compared to text generation.

• Source: Why Sora Failed: $15M/day inference cost vs. $2.1M lifetime revenue