A $20/month user costs OpenAI $65 in compute. AI video is a money furnace
4 hours ago
- #Video Generation
- #AI Economics
- #Tech Failure
- Sora's shutdown due to unsustainable economics: costs significantly exceeded revenue (e.g., $15M/day peak compute vs. $2.1M lifetime revenue).
- High cost structure: generating a 10-second AI video costs ~160x more than equivalent text due to computational demands (e.g., processing 80,000 pixel patches, multiple refinement passes).
- Low user retention: only 1% day-30 retention compared to TikTok's 32%, with users treating it as a novelty without recurring use cases.
- Failed partnerships and ethical issues: Disney's $1 billion deal collapsed upon shutdown; deepfakes and moderation problems damaged reputation.
- Industry-wide unprofitability: no AI video company (e.g., Runway, Pika) has proven net profitability, highlighting structural challenges.
- Future viability depends on cost reduction: need 100-300x cost cuts (e.g., via Rubin GPUs, distillation techniques) to achieve consumer pricing (e.g., $0.01/clip).
- OpenAI's strategic shift: shutdown aligns with pre-IPO focus on profitable areas (e.g., coding agents, robotics simulation) and compute allocation.