OpenAI has decided to definitively close SoraSora, its ambitious AI-powered video generation platform, is being discontinued just months after launching it as a standalone app with aspirations of becoming a global social network. The decision, communicated internally by Sam Altman and publicly confirmed via Sora's official account on X, marks a significant shift in direction for the company behind ChatGPT.
The move not only implies the disappearance of the consumer app, but also the withdrawal of the API and the rest of the video functions supported by the Sora modelincluding those integrated into ChatGPT. Behind this shift, OpenAI asserts that it needs to concentrate computing power and talent in areas of higher strategic priority, especially in enterprise tools, programming, and robotics.
What was Sora and why is its closure so significant?
Sora was born as artificial intelligence model capable of generating video from text, initially presented in 2024. In a second phase, OpenAI turned it into a consumer application with a markedly social focus: a TikTok or Instagram Reels style clip feed where users could create, share and comment on AI-generated videos.
With the launch of the standalone app in September, Sora became the OpenAI's first proprietary application after ChatGPTDuring the first few days, it climbed to the top of the iPhone App Store and reached one million downloads in less than a week, according to data from analytics firms such as Sensor Tower, which also recorded around 600.000 installations in the last month before the closure announcement.
The initial appeal lay in its ability to create hyperrealistic videos Based on text descriptions and the ability to insert users into iconic scenes from popular culture, Sam Altman himself encouraged people on social media to try the system by recreating movies, TV series, and video games, which sparked creativity… and also legal problems.
Over time, the platform integrated an improved version of the model, known unofficially as Sora 2with higher visual quality, better motion control, and options for expanding existing clips. Even so, within OpenAI, the following was already being discussed: extremely high consumption of computational resources facing a demand that, after the initial boom, was not quite consolidating.
Closure announcement: Goodbye to the app, API, and video on ChatGPT
The farewell was staged with a direct message posted on X by the project's official account: “We're saying goodbye to the Sora app”In that same text, the team thanked those who had created and shared videos, and promised more details about the closing dates for both the application and the API, and about how to export and preserve user-generated content.
Internally, Sam Altman conveyed to the staff that OpenAI is going to dismantle all initiatives based on its video modelsThe withdrawal includes the commercial app, the developer version, and the video generation and editing features that had been integrated into ChatGPT as part of the commitment to multimedia content.
For now, the company has not set a public and closed date for the final shutdown of the serviceHowever, the company has insisted that it will announce the schedule in advance so users can download or migrate their creations. At the same time, the company says it is "exploring ways to support the export and preservation" of videos hosted on the platform.
The announcement took even some members of the community by surprise, since Just a few days earlier, OpenAI had published a guide On its official blog about how to use Sora safely and responsibly, there was no indication that such an imminent closure was being considered. Sources cited by media outlets such as The Wall Street Journal and CNN suggest the decision was made within a broader context of internal reorganization and cost pressures.
A change in strategy: less video, more business and robotics
In parallel with the closure of Sora, OpenAI is carrying out a profound repositioning of its product offeringThe stated goal is to reduce the dispersion of consumer initiatives and concentrate resources on productivity tools, programming, and "agent" systems capable of operating autonomously on the user's computer for tasks such as writing code, analyzing data, or automating workflows.
Within that plan lies the integration of the ChatGPT desktop app, the code development tool (derived from Codex), and its browser into a kind of unified “superapp”This merger aims to align the different engineering teams under a single product vision and offer a more consistent experience for businesses and advanced users.
The company itself acknowledges that Computing costs are a determining factor in its decisions. Video generation is one of the most infrastructure-intensive tasks: it requires powerful GPUs, large data centers, and ongoing energy and maintenance costs. In a scenario of growing demand and limited hardware, OpenAI says it is forced to make "compromises" and prioritize those lines with the greatest economic and strategic return.
This change of focus fits with the company's plans to strengthen your business compared to rivals like Anthropic or Google. The former has gained ground with products like Claude Code, highly valued by developers, while Google boasts advances in its own video models and integrates its systems throughout the search engine's ecosystem of services.
Part of the team that worked in Sora will be relocated to robotics and real-world simulation projectsOpenAI wants to leverage video generation technology to model physical environments and train robots in complex tasks before deploying them in the real world. In other words, Sora is no longer a consumer product, but some of its underlying technology will be repurposed as an internal research tool.
Copyright, deepfakes and regulatory pressure
Sora's closure is also better understood if you review the long list of controversies linked to intellectual property and the use of images of real people. From its earliest days, rights holders expressed concern about how easily protected characters, logos, and scenes could be recreated without authorization or financial compensation.
The platform quickly filled with videos that They imitated styles, franchises, and public figuresThis sparked complaints from entertainment companies, video game studios, and other content owners. Some critics pointed out that Sora was fueling a flood of "AI garbage" and facilitating the creation of increasingly believable deepfakes, with the risk of disinformation, defamation, and identity theft.
Faced with that scenario, OpenAI began to introduce blocking and filtering mechanisms so that rights holders could prevent the use of their intellectual property within the app, such as when suspended Martin Luther King Jr. videos on SoraBut those same restrictions detracted from much of the appeal that had driven its initial virality, based precisely on being able to mix icons of popular culture with hardly any limits.
Sora's case illustrates the extent to which The battle over copyright has become a bottleneck for generative AI. It's not just about images and video: it also affects text and audio models trained on large amounts of copyrighted material. Regulatory pressures in the United States and Europe, along with ongoing litigation, are forcing tech companies to proceed with greater caution.
At the same time, the public debate about the risks of deepfakes and audiovisual manipulation has intensified. For a company preparing for a potential IPO and already under scrutiny from regulators and governments, to hold such an exposed product how Sora could become a reputational and legal burden that would be difficult to justify to shareholders and partners.
The agreement with Disney and the pacts that are left up in the air
One of the elements that has attracted the most attention in the sector is the impact of Sora's closure on the announced alliance with The Walt Disney CompanyAt the end of last year, OpenAI announced a landmark agreement that would allow the use of more than 200 characters from Disney, Pixar, Marvel, and Star Wars within the platform, under a formal licensing framework and, according to various reports, linked to a $1.000 billion investment in the startup.
That pact was presented as a possible a way to reconcile generative AI and respect for intellectual propertyInstead of using content without permission, Disney would pay for an official catalog. However, OpenAI's change of direction has put the deal on hold. Disney spokespeople have confirmed that the agreement "will not move forward" and that the company "respects OpenAI's decision to exit the video production business."
Disney's withdrawal is not the only front. Other companies in the entertainment and video game sector, such as leading Japanese studios, had expressed their concerns. reservations or legal actions in response to the unauthorized use of their characters and universes in videos created with Sora, collective pressure has made the "create first, negotiate later" model increasingly unsustainable.
For OpenAI, the closure of Sora means giving up on a potential line of business based on branded licenses and contentBut it also means escaping an environment fraught with legal uncertainties. Disney itself has left the door open to further experimentation with other AI platforms in the future, provided that the rights of creators and IP holders are respected, suggesting that the dialogue between technology and entertainment will continue, albeit through different avenues.
Competition, costs and a changing cycle in generative AI
The competitive context helps explain why OpenAI has opted for a drastic cut. While Sora was trying to establish itself as a mass-market creative platform, Anthropic was strengthening its position in the business market With models like Claude, especially valued for their performance in programming and professional tasks. Market reports indicate that the company led by Dario Amodei has captured a very significant share of the spending by companies starting out with AI solutions.
At the same time, Google has intensified its commitment to multimodal models integrated into its search engine and productivity suite, leveraging its advantage in infrastructure and distributionIn this arena, OpenAI no longer competes solely as a charismatic pioneer, but as a company more obligated to demonstrate sustainability and focus, especially if it aspires to go public in the short term.
Sora symbolized a stage in the company's history characterized by continuous releases and exploration of new formatsFrom general-purpose chatbots to creative tools, the app's closure and the retreat of video point to a different phase, in which it's more important to concentrate efforts on a handful of products with a direct impact on revenue and corporate adoption.
In parallel, the AI industry faces a major infrastructure challengeBuilding data centers capable of handling the computing demands of large-scale AI models is crucial. Video generation, due to its cost, directly competes with the deployment of other, potentially more profitable, AI functions, ranging from office assistants to data analytics systems. Faced with this dilemma, OpenAI has opted to sacrifice Sora to free up resources.
All this reorganization is taking place while the company is also promoting initiatives such as the OpenAI FoundationSam Altman announced a $1.000 billion investment for the company, which is restructuring to operate as a traditional corporation with a non-profit arm. The balance between mission, business, and risk is a key factor in many of its current decisions.
The end of Sora marks the close of a very intense chapter in the relationship between artificial intelligence and generative video. What began as one of the most spectacular demonstrations of AI's creative potential now becomes an example of the extent to which it influences... costs, regulation, copyright and competition When it comes to keeping a product alive, OpenAI is moving away, at least for now, from the race to lead in video for the general public and redirecting its focus towards productivity and robotics, leaving a question hanging in the air that looms over the entire sector: how far can generative creativity be scaled when technology, law, and business collide head-on?
