TLDR: OpenAI’s advanced text-to-video generator, Sora 2, is reportedly facing substantial financial challenges due to higher-than-anticipated user generation rates and the creation of content for niche audiences. This financial strain is exacerbated by controversies surrounding the platform’s misuse, including the generation of deepfakes of deceased public figures and other problematic content, raising questions about the company’s content moderation and long-term sustainability.
OpenAI’s latest text-to-video artificial intelligence model, Sora 2, is encountering significant financial difficulties, as reported by Futurism on October 7, 2025. The core of the problem stems from an unexpected surge in user activity, with individuals generating ‘much more than we expected per user,’ and a considerable volume of videos being produced ‘for very small audiences.’ This high generation rate translates directly into substantial computational costs, primarily for Graphics Processing Units (GPUs), which are essential for running such advanced AI models.
Sam Altman, CEO of OpenAI, has reportedly acknowledged that the current ‘slop’ feed – referring to the vast array of user-generated content – serves as a ‘money grab to fund GPUs.’ This suggests a strategic move to attract a large user base by initially offering the service for free, thereby generating data and engagement, while simultaneously seeking to offset the immense operational expenses. It is anticipated that serious users will eventually need to subscribe to a ChatGPT Pro membership, priced at $200 per month, to access the full capabilities and fewer limitations of Sora 2.
Beyond the financial pressures, Sora 2 has been embroiled in several controversies regarding its content. Despite OpenAI’s stated measures to ‘block depictions of public figures,’ the platform has seen a proliferation of videos featuring deceased celebrities. Examples include disturbing depictions of Stephen Hawking being ‘horribly brutalized,’ Michael Jackson rapping, Tupac Shakur in North Korea, and John F. Kennedy discussing Black Friday deals. OpenAI has clarified that it ‘allow[s] the generation of historical figures,’ which has opened the door to such contentious content. Concerns have also been raised about the creation of nonconsensual deepfakes, copyrighted characters, and other intellectual property without proper authorization. More alarmingly, reports indicate that Sora 2 has been used to generate advertisements for ‘Epstein Island’ children’s toys and realistic videos of people shoplifting.
Also Read:
- Indian Entities Take Legal Stance Against ‘AI Slop’ Amidst Generative Video Boom
- xAI Unveils Grok Imagine v0.9: Rapid Voice-First AI Video Generation Arrives on X
Sora 2, described as ‘genuinely levels ahead of the previous iterations of video generation models,’ was launched as a TikTok-style video app for iOS. OpenAI’s overarching goal with this widespread availability is to achieve ‘virality’ and embed the AI in the ‘conscious mind of the everyday person,’ a step they view as crucial on the path to Artificial General Intelligence (AGI). However, the current financial strain and the ethical dilemmas posed by its content highlight significant challenges for the company as it navigates the complex landscape of advanced AI deployment.
