OpenAI Shuts Down Sora Six Months After Launch, Streamlining Operations Ahead of Planned IPO

OpenAI announced Tuesday that it will wind down Sora, its generative AI video platform, approximately six months after the tool’s public launch. The company will also discontinue the Sora API, which previously gave third-party developers and major Hollywood studios access to its text-to-video AI model.

The move signals a clear strategic shift: the ChatGPT developer is streamlining its sprawling portfolio of projects to focus on core priorities ahead of its planned initial public offering. OpenAI Chief Financial Officer Sarah Friar told CNBC in an interview Tuesday that the company needs to get “ready to operate as a public company.”

Since ChatGPT first launched to the public in 2022, CEO Sam Altman has run OpenAI in a similar mold to Y Combinator, the iconic Silicon Valley startup incubator he previously led. The company has placed exploratory bets on a wide range of experimental products, from Sora and a standalone web browser to a line of hardware devices, humanoid robotics projects, and Codex, OpenAI’s AI-powered coding assistant.

Many of these side projects have delivered uneven results, and Sora’s user growth has flatlined dramatically in recent months. Third-party analytics firm Appfigures data shows that after hitting a global peak of 3.3 million combined downloads across iOS and Android in November 2025, monthly Sora app downloads plummeted to just 1.1 million by February 2026.

Over the past several years, OpenAI researchers have described the company’s internal culture as “bottom-up”: rather than sticking to rigid executive-mandated product roadmaps, the company distributes resources to promising emerging ideas as they arise. While this approach has fostered a productive environment for cutting-edge AI research, it has also stretched the company’s limited GPU computing resources and engineering staff far too thin, according to multiple insider sources.

Now, OpenAI’s leadership has issued a clear mandate to refocus the company’s efforts on just a handful of high-priority core areas.

One of the top priorities is a new “super app” that will merge ChatGPT, Codex, and Atlas into a single unified offering. OpenAI’s leadership expects that combining these tools under one consumer-facing interface will help transform ChatGPT into a fully functional universal super assistant. The Wall Street Journal previously reported on the super app project and OpenAI’s broader push to simplify its overstretched product lineup.

Before OpenAI launched ChatGPT in 2022, the company’s original long-term goal was to build a general AI agent capable of completing nearly any digital task for users. Back then, this product was dubbed the “super assistant,” and it was framed as the first real-world realization of artificial general intelligence (AGI) — but insiders note building the tool has proven far more challenging than OpenAI’s team initially anticipated. The company has since tested incremental agent features within the existing ChatGPT platform, including Operator and ChatGPT Agent, though user adoption of these tools has remained underwhelming. OpenAI now believes a new consumer agent built on top of Codex will resonate more strongly with existing ChatGPT users.

OpenAI is also working to grow its enterprise business segment as it prepares for its public market debut. While Anthropic was long the front-runner in the race for enterprise AI coding tools, OpenAI’s Codex team has closed the gap dramatically over the past year. Codex is now one of OpenAI’s brightest spots, hitting $1 billion in annualized revenue in January 2026 and continuing to grow at a steady clip.

Though Sora launched to widespread media hype and industry excitement, the product did not align with OpenAI’s new streamlined strategic direction, leading leadership to conclude the company’s GPUs and research talent would deliver more value allocated to other priorities. In a statement provided to WIRED, an OpenAI spokesperson explained: “As we focus our efforts and demand for computing resources grows, the Sora research team will shift to work on world simulation research to advance robotics that will help people solve real, physical-world tasks.”

The decision to shut down Sora appears to have derailed OpenAI’s high-profile partnership with The Walt Disney Company, which had previously committed to a $1 billion investment in OpenAI. Multiple reports indicate Disney was caught completely off-guard by the Sora shutdown announcement, and the entertainment giant now says it has scrapped its planned investment.

It remains unclear what OpenAI’s new era of strategic focus will mean for the company’s broader research divisions. OpenAI is already locked in fierce competition with Anthropic, Google DeepMind, and Meta for a limited pool of top-tier AI research talent. In January, OpenAI Vice President of Research Jerry Tworek left the company after failing to secure sufficient resources for his next major research project. While many current OpenAI employees have reacted positively to the refocusing mandate, other staff working on deprioritized side projects may choose to leave for competing AI labs, industry observers note.