If someone told me a few years ago that OpenAI would actively shut down its most popular consumer product, I would probably think it was a joke.

After all, when Sora was released at the end of 2024, that sense of shock was real. The videos that "inserted oneself into movie scenes" spread on social media faster than any product launch. Within less than 5 days, downloads exceeded 1 million.

Some called it the “iPhone moment of the AI era.”

However, on March 24 local time, OpenAI announced it would shut down the consumer application of Sora, the developer version, and the embedded video feature in ChatGPT.

The voices that once applauded Sora are probably now all asking the same question:

Why is OpenAI cutting Sora?

The decline of Sora is actually much faster than most people realize.

By December 2025, app downloads dropped by 32%. In January 2026, installations further declined by 45%, with an absolute number still at 1.2 million, but the trend said it all:

Users stayed, but enthusiasm did not.

The reason is not complicated. The feature of "inserting oneself into movie scenes" is essentially a good demo, but not a repeatedly usable scenario. Most people tried it once, found it amazing, and then didn’t know what else to do with it.

The more realistic problem is that many users are unwilling to hand their likeness over to an AI application. This is not a technical problem, but a trust issue. In the absence of celebrity faces and brand support, the content Sora could generate quickly hit the ceiling of "interesting but useless."

The content licensing partnerships that were once highly anticipated are also quietly loosening.

At the same time, both Variety and Hollywood Reporter reported: Disney announced it was terminating its partnership withOpenAI, along with the originally agreed $1 billion investment plan.

According to the original concept, Sora was supposed to generate videos from over 200 characters from Disney, Marvel, Pixar, and Star Wars, opening up "fan inspiration" creation to users earlier this year through ChatGPT.

It is said that a team from Disney was working alongside Sora's engineers the night before, and woke up the next morning to find out the partnership was terminated. That kind of unexpected situation is probably not less than what the users experienced.

If the user loss is the superficial reason for Sora's shutdown, then the IPO is the true behind-the-scenes driver.

OpenAI just completed a new round of financing at $11 billion, pushing the company's valuation to $730 billion. At this juncture, every allocation of computing resources must be scrutinized by investors and potential public market shareholders.

Sora is an extremely resource-intensive product. Generating video is much more expensive than generating text and also much more than generating images.Every time a user "plays" an AI video with Sora, the GPUcomputing power burned behind it could answer dozens of questions on ChatGPT.

When OpenAI's application head Fidji Simo clearly stated in an all-hands meeting that the company needs to stop being distracted by "side quests" and aggressively shift toward coding and enterprise users—Sora's fate was basically sealed.

This was not a difficult decision but a clear-headed one.

Redirecting computing resources away from Sora to Codex (OpenAI's AI programming assistant) makes immense commercial sense.

Codex has seen user growth of threefold this year, usage increase of fivefold, and weekly active users have surpassed 2 million. Last month, OpenAI also acquired the developer tool company Astral, and the entire team was directly integrated into the Codex team.

The direction has been set.

On the other side, Anthropic provided a more straightforward statistic: annual revenue exceeded $19 billion, of which about 80% came from enterprise clients. CEO Dario Amodei mentioned that in just February, they added $6 billion in revenue, almost entirely from Claude Code.

Enterprise clients will pay, will renew, and will expand usage. Consumer users may like, may share, but do not necessarily pay. This is the most painful yet crucial lesson for the AI industry from 2025 to 2026.

The answer may be negative. Or at least, OpenAI chose not to wade through this murky water.

In fact, competition in AI video generation continues and even intensifies. ByteDance's Seedance 2.0 is still active, and Google DeepMind's Veo 3 is also reaching out to filmmakers.

Sora's exit from the consumer market does not imply that this direction itself is unfeasible.

However, a structural problem Sora faced is one that competitors cannot easily bypass, which is the issue of “AI trash.”

When a tool can easily generate hyper-realistic videos, a flood of low-quality AI content begins to emerge on social media. This is not just a regulatory difficulty for platforms; it quietly erodes users' overall trust in AI-generated content. When "AI manufacture" becomes synonymous with low-quality content, the brand value of an AI video application will be severely dragged down.

Paul Roetzer, founder of Marketing AI Institute, has a poignant comment: he believes that the underlying video generation technology of Sora is "incredible," but OpenAI chose to make it into an endlessly consumable entertainment tool, "completely opposite to the direction these laboratories should take."

This judgment somewhat explains why, although technically leading, one can still fail commercially.

Sora is not lacking technology; it lacks an answer to "why would users continue to use it.".

In consumer entertainment scenarios, this answer remains vague; but in fields like professional creation, film industry, and advertising production, the answer may be much clearer.

However, OpenAI does not plan to search for this answer by itself at this time. At least not in the form of the Sora consumer product.

Looking back, OpenAI has indeed been somewhat scattered in its product line over the past year or so.

ChatGPT, Sora, image generation, voice assistants, enterprise APIs, Codex, custom GPT… each direction intended to be developed, each achieving a "good" result, but "good" is becoming less valuable in this industry.

Now, the story has changed.

Sora going offline is a signal and a choice.

OpenAI tells everyone: we know what can make money, and we will focus resources there. At the same time, the plan to expand the workforce from 4,500 to 8,000 employees by the end of the year indicates that the company is not shrinking but concentrating efforts on more important things.

This company is taking action to answer a question many AI companies are still evading: are you demonstrating technology, or are you building a real business?

Sora's fireworks were indeed beautiful.

But once the fireworks fade, what remains is what the company truly desires.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。