OpenAI’s shutdown of Sora was not framed in the source as a mystery about technology limits or a public-relations retreat. The simpler explanation is that the AI video-generation tool became too expensive to keep running at the scale it had, while its audience was shrinking.
Sora had been public for just six months. It allowed users to upload their own faces and place themselves inside generated video scenes, which immediately led to suspicion after the shutdown. But according to a new WSJ investigation cited in the source, the real issue was far less dramatic: the app was costing too much, too few people were using it, and OpenAI needed the resources elsewhere.
Usage fell while costs stayed large
Sora began with attention and momentum, but that early interest did not hold. The source says its worldwide user count peaked at around a million and then fell to fewer than 500,000.
For many digital products, that kind of drop would be a concern. For an AI video-generation tool, it becomes a deeper operational problem because every use is tied to expensive computing resources. The source says Sora was burning through roughly $1 million every day.
That cost was not presented as a sign that people were heavily committed to the product. Instead, it reflected the expense of video generation itself. Creating AI video requires significant compute, and the source describes every user who placed themselves into a generated scene as drawing from a finite supply of AI chips.
The business problem was therefore direct: Sora was expensive even as its user base was contracting. The more OpenAI kept the app alive, the more it had to keep assigning scarce resources to a product that was not, based on the source, delivering enough usage to justify the cost.
The compute tradeoff became impossible to ignore
The Sora shutdown also shows how AI companies face a different kind of product decision from many software firms. Keeping a feature online is not only a question of engineering maintenance or customer support. For tools like Sora, it can mean committing large amounts of compute every day.
In the source’s account, OpenAI was not simply deciding whether Sora was interesting. It was deciding whether Sora deserved chips, people, and attention at a moment when those same resources could support other priorities.
That is why the decision matters beyond Sora itself. The source says a whole team inside OpenAI was focused on making Sora work. At the same time, Anthropic was gaining ground with software engineers and enterprises, which the source identifies as the customers that drive revenue.
Claude Code is singled out as an especially important competitive pressure. The source describes it as taking business from OpenAI, and that competitive context made Sora’s resource needs harder to defend.
Sam Altman chose refocus over persistence
According to the source, CEO Sam Altman made the decision to shut Sora down, free up compute, and refocus. That framing is important because it presents the shutdown as a strategic reallocation rather than a narrow product cancellation.
The logic was straightforward. If Sora required a large daily spend, depended on finite AI chips, and had fewer users than it once did, then keeping it running meant accepting an opportunity cost. The resources used to operate and improve Sora could not be used elsewhere at the same time.
For OpenAI, the pressure was not only internal. The source says Anthropic was quietly winning over the software engineers and enterprises that matter to revenue. In that environment, an expensive consumer-facing video app with declining usage became harder to prioritize.
The decision also underlines a practical reality for AI products: attention at launch is not the same as durable demand. Sora attracted enough interest to peak at around a million users, but the later drop to fewer than 500,000 changed the equation. A product can be impressive and still be a poor use of resources if the cost of each interaction is too high.
The Disney deal ended with the shutdown
The speed of the decision is clearest in what happened to Disney. According to the source, Disney had committed $1 billion to the partnership. Yet it found out Sora was being shut down less than an hour before the public.
The source says the deal died with Sora. That detail shows how abrupt the move was and how far the consequences reached beyond OpenAI’s own product lineup.
For Disney, the shutdown meant a major partnership disappeared at the same time as the product behind it. For OpenAI, it showed that the company was willing to end a high-profile arrangement if the underlying product no longer fit its priorities.
What Sora’s shutdown says about AI products
The Sora story is ultimately about discipline in a compute-constrained market. The source does not present the shutdown as a rejection of AI video in general. It presents it as a decision about cost, usage, and competitive focus.
Several lessons follow directly from the facts given:
- Launch attention is not enough. Sora’s user count peaked at around a million, then fell to fewer than 500,000.
- Video generation carries heavy operating costs. The source says Sora was costing roughly $1 million every day.
- Compute is a strategic asset. Every Sora user consumed part of a finite supply of AI chips.
- Enterprise demand shaped the decision. Anthropic was gaining traction with software engineers and enterprises while OpenAI’s team was still trying to make Sora work.
- Partnerships depend on product survival. Disney’s $1 billion commitment ended when the Sora deal died.
Seen that way, OpenAI’s move was not just about one app. It was about choosing where to spend scarce compute, which customers to prioritize, and how quickly to abandon a product that no longer matched the company’s most urgent competitive needs.