Today’s “Hollywood Horror Show” is a reflection of a scenario shaped by Artificial Intelligence, with high-tech development and growing consumer acceptance. As Doug Shapiro pointed out just over a year ago, the value of producing videos is quickly converging with the cost of computing.
In the near future, the below-the-line price (not counting talent) of a blockbuster quality film will drop dramatically from $1-2 million per minute to just $10-20 per minute. This creates a nearly endless supply of high-value, low-production content.
Shapiro’s reflections become even more relevant in light of the recent collapse of OpenAI’s Sora. The project, which promised to revolutionize the creation of videos with AI, was unable to overcome the economic barrier and was closed last week.
Continues after advertising
Disney terminated a US$1 billion contract and, although the application was launched with great expectations, it ended its journey without its 200 licensed characters having even been used by consumers.
In a long essay published in February last year, Shapiro wrote that the last 10-15 years of video have been dominated by disruption in content distribution. However, he suggested that the next ten years would be defined by the disruption in content creation.
Sora’s failure exposes a more complex problem: the promise of democratizing content comes up against the economics of UGC and the risk of a business model that has not yet proven itself.
Continues after advertising
The content creation landscape and the IP economy
OpenAI’s decision highlights a point that Ben Woods has insisted on: the value in the content creation chain has not yet been established, especially when it comes to licensing intellectual property (IP) and fan-generated content.
The analyst argues that the value of UGC remains underexploited in traditional monetization models. For him, the true transformation of the creator economy will not come from the pure automation of creation, but from the intelligent integration between UGC and licensed IP. He warns that, without this combination, the value of user-generated content will be difficult to achieve in a sustainable way.
IP monetization cannot be seen just as a matter of access to large audiences or volume of content. The sustainable business model, as pointed out by Woods, involves creators participating in the value generated, with smart licensing and co-ownership of IP, which would give creators a greater share of the revenue generated by their work, rather than just a one-way distribution.
Continues after advertising
“My conclusion is that it was mainly a cost decision, and the platform and user base were never seriously developed. So we still have to wait and see about AI and fan creation,” he argues.
The account did not close
The economic point highlighted by Woods is reinforced by the data revealed by Forbes. At its peak, Sora was costing OpenAI around $15 million per day, which represented approximately $1.30 per video generated.
However, revenue from in-app purchases was only $2 million. This value barely covered three hours of operation, as ProfgMedia highlighted
Continues after advertising
A practical example given by the Entertainment Strategy Guy is the Coca-Cola Christmas commercial, produced with AI. The headlines indicated that it took 100 people to generate 70,000 clips. At a cost of $5 per clip, that works out to a total of $350,000 for the commercial.
Although the real value has not been disclosed, the case indicates that production at scale with AI is still expensive, and a large part of the expenses are subsidized by technology companies trying to capture the market before raising prices.
Compared to other large-scale projects, such as Horizon Worlds, Facebook’s metaverse, which had a peak of 300,000 monthly active users, and MySpace, which has less than 800,000 monthly visitors today, Sora was far from reaching the scalability necessary to be considered a viable model.
Continues after advertising
As ProfgMedia wrote, there was no possibility of such a disastrous parallel investment continuing.
The 1% Rule and the Creation Misconception
If the bill didn’t add up from a cost point of view, it also didn’t hold up on the demand side.
According to data from Olivia Moore of Andreessen Horowitz, Sora’s first-day user retention was 10%. On the seventh day, it dropped to 2%. In the thirtieth, 1%. In the sixtieth, close to zero.
Analyst Jen Topping uses the “1% rule” to put it into context: around 1% of users create content, 9% participate and 90% just consume it. In practice, the concentration is even greater. As an example, Topping cites Instagram with its 2 billion users, but remembers that hardly ten million (or 0.5%) can be considered relevant creators.
Sora’s implicit bet was to reverse this logic. Make creation accessible enough to turn consumers into creators. And, from this, generate a new layer of content capable of sustaining engagement.
It didn’t work.
When the novelty effect disappeared (both for those who created and for those who watched) the behavior returned to the original pattern. Technology reduced the cost of production, and created no reason to do so.
This point becomes even more evident when we broaden our view of consumption. As ProfgMedia highlighted, the 1% rule can only show holes on the surface.
The actual distribution of attention is even more concentrated. See the numbers shared by the expert: 4% of YouTube videos account for 94% of views 3% on Instagram account for 84% 5% on TikTok generate 89% And the 25 largest podcasts capture almost half of the weekly audience in the US
The videos generated by the machine that went viral on the networks generated the perception of rupture. In the end, do people want to create or just consume better? ProfgMedia argues that consumers do not want to produce their own entertainment: “They want to be entertained.”
And for now, the answer lies in retention.
