OpenAI has indefinitely shelved its planned erotic chatbot feature, citing safety concerns and investor pressure, as the company accelerates a sweeping consolidation of its product line ahead of a possible 2026 public listing worth up to $1 trillion.
SAN FRANCISCO — OpenAI indefinitely paused its planned “adult mode” for ChatGPT on March 25, 2026, according to a report by the Financial Times, ending months of internal debate over a feature that had drawn unanimous opposition from the company’s own safety advisers. The move is one of several rapid product cuts at the company, which is now explicitly orienting itself toward enterprise software clients ahead of what could be the largest AI IPO in history.
The decision was not just about content risk. It came alongside the March 24 shutdown of Sora, OpenAI’s AI video platform — killed less than six months after launch — and the collapse of a $1 billion content deal with Disney, which had been counting on Sora to generate content using characters from Marvel and Star Wars. The pattern is clear: consumer experiments that drain compute or carry reputational exposure are being cut.
The age-verification problem that killed the feature
What has gone largely underreported is that the erotic chatbot’s cancellation was not primarily a moral decision — it was a technical one that exposed a serious legal liability. OpenAI’s own age-prediction system, built specifically to gate adult content, misidentified minors as adults at an error rate exceeding 10 percent, according to the Wall Street Journal. With 800 million weekly active users, even a 10 percent failure rate would expose millions of underage users to explicit content — an insurmountable compliance risk for a company actively courting institutional investors.
OpenAI launched the age-prediction tool in January 2026, framing it as a safety feature that would automatically restrict content for users under 18. Adults could override it via a selfie submitted through identity-verification service Persona. But the underlying model was not accurate enough to rely on for content this sensitive, and internal teams apparently knew it.
OpenAI’s advisory council on AI and well-being, comprising experts in psychology and cognitive neuroscience, unanimously warned the company in January not to proceed. One adviser warned the company risked creating what they called a “sexy suicide coach” — a phrase that captures why investors and employees raised objections well before the Financial Times reported the halt. The council’s opposition, combined with the technical gap in age verification, made the feature legally and commercially untenable.
A superapp strategy replaces the consumer sprawl
Rather than building new standalone products, OpenAI is now merging ChatGPT, its Codex coding agent, and its Atlas browser into a single desktop application. Fidji Simo, OpenAI’s applications CEO, confirmed the consolidation in a company-wide meeting, calling existing fragmentation a barrier to quality. CEO Sam Altman and chief research officer Mark Chen have both identified “side quests” — consumer features outside the core enterprise use case — as products the company can no longer afford to maintain.
The financial reality sharpening these decisions is stark. OpenAI is reportedly facing potential losses exceeding $14 billion this year, even as it reports $25 billion in annualized revenue and prepares a pre-IPO roadshow targeting a valuation of up to $1 trillion, according to documents shared with prospective investors in March 2026. The company raised more than $100 billion in its latest funding round at a $730 billion valuation, and is offering private equity investors returns of up to 17.5 percent alongside early model access to attract further capital.
That pressure explains the abruptness of the Sora shutdown in a way that computing costs alone do not. Disney’s investment had been confirmed just three months earlier. The speed of the reversal suggests the IPO calculus — not just server costs — drove the timeline.
What the adult mode cancellation ultimately reveals is a company that spent much of 2025 testing the outer limits of what a general-purpose AI could do, and is now retreating to a narrower, more defensible business story for Wall Street: enterprise coding tools, agentic AI systems, and a unified platform that competes directly with Microsoft Copilot and Anthropic’s Claude for corporate contracts. Adult content, video generation, and AI shopping features don’t fit that pitch. They’re gone.
