Over the past year, OpenAI has tried to span as wide a range of businesses as possible—from video services and embedded commerce to experiments with the generation of erotic content. Now the company is rapidly shifting its focus toward what can deliver durable revenue.
That means abandoning its riskiest consumer-facing features—in particular, adult content—while intensifying work on corporate products and monetisation, especially as competition from Anthropic grows. Business customers now offer the clearest and most scalable revenue models: they want tools for generating text and building entire systems of agents capable of multiplying employee productivity, not entertainment-driven forms of interaction with chatbots.
The idea of introducing erotic content was already under discussion in October, when OpenAI chief executive Sam Altman raised the possibility against the backdrop of reports that ChatGPT user engagement was slipping. But during testing, the project ran into technical difficulties—in particular, there were problems removing unwanted references to bestiality and incest, the Financial Times reported.
This is already the third decision in recent days to pull back from consumer products. OpenAI has also shut down Sora, the video service that gained broad traction after its September launch, and scaled back Instant Checkout—its built-in shopping tool—abandoning the idea of processing transactions directly.
An additional source of pressure came from Apple’s decision, first reported by Bloomberg: the company plans to allow third-party chatbots, including Claude and Gemini, to integrate with Siri, intensifying the competitive environment.
OpenAI was originally founded as a research lab, and it therefore pursued a wide range of initiatives in parallel—from a web browser and music generation to wearables, smart speakers and glasses. The company now says it intends to concentrate on solutions that are already working, stop spreading its resources too thin and make use of the present moment.
At the same time, OpenAI is not abandoning the consumer segment—and users, despite their interest in Claude, are not leaving either. ChatGPT still has 900 million weekly active users and 50 million paying subscribers.
The strategy now taking shape points to preparations for a public listing and an effort to turn that vast user base into a durable stream of income. “ChatGPT is the entry point to the world of AI for most users,” the company said in a blog post announcing the raising of $110 billion in investment to “expand access to advanced AI technologies for people, businesses, and communities around the world.”
On Tuesday—the same day OpenAI announced Sora’s shutdown—the company published an updated version of its Model Spec, a 100-page document governing ChatGPT’s behaviour, similar to the ones Anthropic regularly updates for Claude. In it, the company’s mission is framed around the “democratisation of access” to AI in healthcare, science, education and work—with no emphasis on entertainment features.
Judging by the reaction, OpenAI’s retreat from the more provocative uses of ChatGPT suits most observers. “Public pressure over AI’s impact on child safety and mental health has been steadily building ever since OpenAI first announced plans to allow adult content,” said Jessica Ji, a senior analyst at the Center for Security and Emerging Technology.
“I would view this first and foremost as a purely commercial decision—similar to the shutdown of the Sora video-generation project,” said Carey Gus Hurwitz, senior fellow and academic director of the Center for Technology Innovation & Competition at the University of Pennsylvania Law School. In his words, “this is not the moment when the company is prepared to alarm potential investors with legal, regulatory or political uncertainty.”