[DigitalToday reporter Jinju Hong] With OpenAI deciding to end the video-generation app Sora service in March 2026, attention is focused on why it is stopping operations.

On March 31 (local time), online media outlet Gigazine cited a Wall Street Journal report as saying Sora was pushed down the business priority list after failing to deliver clear revenue despite high operating costs.

Sora is a free app that generates videos from text input. It also offers a feature that lets users or acquaintances appear in videos, drawing major attention early on after its launch. Despite an invitation-only system, it quickly attracted users, including reaching the top ranks on the app store. As it spread, controversies over copyright and deepfakes emerged. Issues were repeatedly raised, including cases in which large numbers of deepfake videos related to Martin Luther King Jr were created and bereaved family members protested, as well as problems of copyrighted content being generated without permission.

In response, OpenAI promised improvements while exploring consent-based content-use models, including a contract with The Walt Disney Company, but it did not overcome fundamental business viability problems. According to WSJ, Sora turned into a financial burden for the company within months of its launch. Its position narrowed as the company strengthened a strategy of focusing on more profitable services ahead of an initial public offering (IPO).

User metrics also declined. According to remarks by a person concerned, Sora’s user base fell from about 1 million soon after launch to below 500,000. At the same time, it was reported to have generated losses of about $1 million a day, or about 1.5 billion won. It was an analysis that sustainability weakened as a high computing-cost structure combined with a low level of monetisation.

Resource reallocation is also cited as a key background to the shutdown decision. OpenAI is in the final stages of developing a next-generation AI model codenamed Spud. It is known to be allocating more computing resources to coding and enterprise product lines, including that model.

Going forward, OpenAI plans to focus on agent-type AI and a super app strategy. That is a direction of integrating AI tools that autonomously perform various tasks, such as software development, data analysis and travel bookings, into a single platform.

Organisational change is also being detected. Sam Altman (샘 올트먼), OpenAI’s chief executive officer, was reported to have instructed the Sora development team to focus in the future on long-term research areas such as robotics.

The decision shows the generative AI industry is being reshaped around profitability and cost structures, moving beyond a stage of simple technology demonstrations. It is evaluated as a case that shows compute-intensive services such as video generation may struggle to survive without a clear revenue model and a system to respond to regulation. At the same time, it suggests OpenAI is shifting its centre of gravity during listing preparations from experimental consumer-facing services to enterprise solutions and long-term technology development.