OpenAI Shuts Down Sora to Free Up GPU Capacity

OpenAI has shut down Sora, its text-to-video product, and moved the underlying GPU capacity to other priorities within the company. The Wall Street Journal reported the decision, noting that Sora had been running as a separate operation from OpenAI's main research effort before the company pulled the plug. The numbers tell part of the story. Sora briefly hit the top of Apple's App Store after launch, drawing significant consumer attention. But attention does not pay for GPU time. The product required sustained compute investment, and OpenAI appears to have decided the return did not justify the cost, especially with the company racing to ship and scale its core language models. This is one of the clearest recent examples of how expensive AI product strategy has become. At companies like OpenAI, product lines compete not just for users but for compute, margin, and internal relevance. A flashy consumer feature with limited monetization becomes a liability when GPUs are scarce and the core platform needs feeding. For developers and businesses building on top of AI products, this is worth watching. A tool you depend on can disappear not because it failed but because it lost an internal resource allocation fight. That is a different kind of platform risk than the usual API deprecation cycle. For Nigeria's growing class of AI-powered startups and builders, the lesson applies directly. Choosing infrastructure and tooling from providers with strong commercial fundamentals reduces the risk of sudden shutdowns. OpenAI's move also signals that video generation, at current economics, is not yet a sustainable product category for the biggest AI labs. Building on AI requires understanding not just capability but the business logic behind the tools you use.