On March 24, OpenAI announced the shutdown of its Sora app, a move that has raised eyebrows in the tech community. OpenAI CEO Sam Altman declared that both consumer and developer versions of Sora would be discontinued, along with the video functionality in ChatGPT. Launched in September 2025, Sora quickly gained traction, described by industry insiders as the “GPT moment” for video generation, but its decline occurred alarmingly fast, leading many to question the rationale behind its abrupt termination.
Technical limitations were a primary factor behind Sora’s downfall. According to Song Chunyu, a senior partner at Lenovo Capital and Incubator Group, Sora’s technical framework was not fully developed. “OpenAI’s Sora is like giving everyone a demo,” he remarked, suggesting that the model still required significant breakthroughs to become viable. Xiao Shi, an AI video entrepreneur, echoed this sentiment, stating that while Sora’s generation capabilities were impressive, it failed to meet essential commercial standards such as high controllability and reproducibility.
Market data further illustrates Sora’s challenges. Although the standalone app garnered over one million downloads within its first ten days, interest waned quickly. By December 2025, downloads plummeted by 32% month-on-month, followed by a staggering 45% drop in January 2026. User spending also exhibited a downward trend, and retention rates were alarmingly low, with just 10% of users returning after one day, and only 0% after 60 days, as disclosed by Olivia Moore, a partner at a16z.
Cost pressures contributed significantly to Sora’s demise. Reports indicated that the app’s monthly computing power costs soared to $15 million, straining the resources of other core teams at OpenAI. Internal communications from Altman highlighted the shutdown as a strategic move aimed at reallocating resources towards enterprise-level productivity tools, especially as OpenAI prepares for an initial public offering (IPO) in the second half of 2026. Ending the unprofitable Sora business would serve to enhance financial performance and present a more favorable outlook to potential investors.
Accompanying the app’s shutdown was the termination of a three-year partnership with Disney, which had involved a $1 billion investment and licensing of over 200 intellectual properties. An investor commented on the broader implications, noting that AI video generation products are still in a transitional phase, experiencing bubbles and shakeouts. “The shutdown process itself is normal,” he remarked, underscoring the volatility inherent in the industry.
The question now looms: does Sora’s exit signify a turning point in the video streaming industry? Just as ChatGPT spurred a surge of large language model startups, Sora had similarly accelerated developments in textual video models. Prior to Sora’s release, companies like AiShi Technology had already introduced platforms such as PixVerse V1 and Runway Gen1, with new entrants continuing to emerge. However, with Sora’s abrupt departure, a reshuffling in the competitive landscape appears imminent.
Industry insiders suggest that this shift favors larger companies that can leverage existing infrastructures to create closed ecosystems, effectively using AI video as foundational technology for their core businesses. This advantage is something startups struggle to match, potentially impacting their commercialization speed. Moreover, fundamental technical challenges remain unresolved across the sector, as indicated by a Conch AI employee who noted common issues with generation stability. “If the generation time is too long, the possibility of crashing later is very high,” he explained.
Copyright compliance presents another hurdle for companies in the space. MiniMax has faced legal challenges and has had to navigate partnerships with film and television institutions to mitigate these risks. Mei Tao, founder of Zhixiang Future, highlighted that high-quality copyrighted data assets will become essential for AI companies. “By 2028, there is reason to believe that large models will consume the existing, ready-made data generated by humans,” he stated, emphasizing the need for strategic planning in anticipation of potential data shortages.
While opportunities are clearer for larger entities, startups still have avenues to innovate. “Startups need to build their own unique interactions or content,” one investor noted, highlighting the undefined nature of the video agent market. This suggests that despite Sora’s failure, the race for speed and innovation remains open, with potential for new entrants to carve out niches in the evolving landscape of AI-driven video generation.
See also
Sam Altman Praises ChatGPT for Improved Em Dash Handling
AI Country Song Fails to Top Billboard Chart Amid Viral Buzz
GPT-5.1 and Claude 4.5 Sonnet Personality Showdown: A Comprehensive Test
Rethink Your Presentations with OnlyOffice: A Free PowerPoint Alternative
OpenAI Enhances ChatGPT with Em-Dash Personalization Feature
