SpaceX warns IPO investors of Grok's 'Spicy' mode risks

SpaceX officially warned potential investors that Grok's 'Spicy' and 'Unhinged' modes – features that allow the chatbot to generate risqué images and voice responses with minimal safety guardrails – could expose the company to regulatory scrutiny and reputational damage. The warning came in a filing Wednesday as part of SpaceX's planned initial public offering. According to the document, SpaceX has already reserved $530 million for potential litigation losses, some stemming from ongoing complaints about Grok generating sexualized imagery of apparent minors. The company is under investigation in the US and other countries over these allegations. SpaceX also notes that future misuse of its AI products, especially the unfiltered modes, could lead to loss of market access, intellectual property disputes, and harmful content generation. While these risks are standard IPO disclosures, they highlight the tension between xAI's stated mission of 'truth-seeking AI' and the practical consequences of releasing features with few filters.

The AI unit, which includes X and xAI, remains a financial drag on SpaceX, posting a $6.3 billion operating loss in 2025 despite $3.2 billion in revenue. Advertising on X dropped $100 million in Q1 2025, though subscriptions to Grok and X grew by $177 million in the first three months alone. SpaceX reported 6.3 million active paid subscribers (4.4 million for X, 1.9 million for Grok) and 550 million combined monthly users (117 million using Grok AI). A bright spot: a deal with Anthropic worth $15 billion annually for data center access. Investors must weigh Grok's user growth and free-speech appeal against mounting legal and regulatory exposure, especially as nonprofits warn that xAI's poor safety record could become a liability for the entire SpaceX enterprise.