Why OpenAI Pulled the Plug on Sora

Video generation tools like Sora promised to change how businesses create content, but OpenAI’s recent decision to shut it down shows the challenges in making such technology practical. Last week, the company ended the app just six months after its public launch. This move highlights how AI services are evolving from flashy experiments to tools that deliver real value for companies and professionals. 

Let’s step back and look at the bigger picture of AI services today. Businesses once dreamed of AI handling everything from writing reports to producing marketing videos. Early tools focused on creative outputs, like turning text into short clips. These caught attention because they felt magical. You could type a description, upload a photo, and watch an AI create a scene with your face in it. For a marketing team, that sounded ideal: quick videos without hiring filmmakers. Yet, as OpenAI learned with Sora, the reality involves heavy costs and limited demand. The company hoped Sora would attract millions of users who needed video content for ads or social media. Instead, interest faded fast after the initial buzz. 

Now consider why video generation turned out to be such a burden. Each video clip required massive computing power. OpenAI reported spending about $1 million a day to keep Sora running. Users peaked at around one million worldwide, then dropped below 500,000. That meant fewer people were justifying the expense. Every time someone generated a clip, it used up specialized AI chips in short supply. These chips power not just videos but also coding tools and data analysis that businesses rely on more. OpenAI had a team dedicated to Sora, pulling resources from other projects. Meanwhile, rivals like Anthropic gained ground with practical alternatives. 

This brings us to how AI services are evolving toward what businesses actually need. Take coding assistance as a brief example. Anthropic’s Claude Code has become popular among software engineers and enterprises. Engineers at top AI firms now use it to write 100% of their code in some cases. Unlike video tools, coding AI helps build products, automate tasks, and speed up development. Companies pay for these services because they save time and reduce errors. OpenAI saw Anthropic pulling ahead in this area, where revenue comes from subscriptions and enterprise deals. Sora, by contrast, appealed more to casual users than paying customers. Shutting it down freed up compute resources to compete in coding and other high-value areas. 

The decision came from CEO Sam Altman, who prioritized long-term goals over a struggling product. OpenAI refocused on areas like advanced language models and developer tools. This shift makes sense in a competitive landscape. AI firms face pressure to allocate chips wisely, as demand outstrips supply. Video generation, while impressive, does not yet match the immediate needs of most businesses. Enterprises want AI that integrates into workflows, such as generating code or analyzing data. Sora’s end signals a broader trend: AI services must prove their worth in dollars saved or earned.

One notable casualty was a planned partnership with Disney. The entertainment company had pledged $1 billion to use Sora for content creation. Disney learned of the shutdown less than an hour before the public announcement. That abrupt end killed the deal and raised questions about communication in big tech partnerships. For Disney, it meant rethinking AI video plans at a time when streaming competition is fierce. OpenAI’s choice underscores that even major commitments cannot save a product without users or profitability. 

OpenAI remains a private company, so it faces no public shareholder pressure. Still, investors and partners expect efficient use of capital. Sora’s failure serves as a case study for other AI firms. Launching consumer-facing tools is exciting, but sustaining them requires steady demand. As AI evolves, expect more focus on backend services like coding and analytics. These deliver consistent revenue and appeal to the enterprises driving growth. Video tools may return in more efficient forms, perhaps integrated into broader platforms. For now, OpenAI’s pivot shows the maturity of the industry: hype gives way to hard economics.

AI promises transformation, but not every idea scales. Tools like Claude Code demonstrate where value lies today. They help teams work faster without draining resources. Sora’s story reminds us that innovation must align with market realities. Companies exploring AI should prioritize needs over novelty to avoid similar fates.