The filing said 117 million use Grok’s AI features monthly but acknowledged the reputational risks posed by the chatbot’s “Spicy” Imagine Mode and “Unhinged” Voice Mode.

SpaceX listed the chance that Grok could generate "potential nonconsensual or exploitative imagery" as a risk factor in its pre-IPO paperwork.

The filing said 117 million use Grok’s AI features monthly but acknowledged the reputational risks posed by the chatbot’s “Spicy” Imagine Mode and “Unhinged” Voice Mode.