A new report in The Information, citing two former x AI employees, says well over half of Grok’s usage stems from NSFW activities. Users generate porn, engage in adult role‑play chats, and request large volumes of erotica, and has drawn scrutiny from regulators and privacy advocates.

Internal data shows the coding model, cheaper to run, receives a “significant proportion” of porn or nude‑image requests, turning adult content into a revenue stream. Although the IPO paperwork omitted NSFW earnings, x AI warned investors that the irreverent feature set posed legal risk and set aside $530 million for potential costs, highlighting its unexpected profitability.

Engineers wrestled with allowing R‑rated chats while blocking child sexual‑abuse prompts, a problem with no quick fix. Staff assigned to “Ani,” Grok’s anime‑styled avatar, reported embarrassment after the bot produced sexualized images of real people, including minors. Despite limits on image edits for paid users, the controversy raises compliance concerns and threatens x AI’s government contracts, where tolerance for NSFW content is minimal.