Sora 2 has just introduced revenue sharing for rightsholders.

For the longest time, the approach that major AI Labs have taken regarding data is: “scrape first, apologize later.”

When Sora 2 was released, there were hundreds of thousands of videos being generated with copyrighted characters from nearly every major franchise.

But due to immense backlash, they are now introducing an opt-in revenue share for copyright holders.

On the surface, it looks like a goodwill move. But what it really signals is something much deeper: even OpenAI is starting to recognize that they’re not above the law.

When the world’s most powerful model begins compensating rightsholders, it means the era of “scrape first, apologize later” is ending.

There are hundreds of these AI scrollable apps coming from Meta, Google, and more in 2025-2026, meaning demand for licensing is now at an all time high, and a successfully positioned company to cleanly arbitrage and source this data will become a unicorn.

The thesis here is that Kled will penetrate the market through a novel concept of "upload your camera roll and get paid" and be the first to 1m users. And simultaneously adapt our product to collect more specific datasets from these users that can move these AI models towards a cleaner and legally secure future.