SpaceX warned investors in an IPO filing submitted Wednesday that AI features such as Grok’s “Spicy” and “Unhinged” modes could expose the company to regulatory scrutiny and reputational damages. These specific modes allow the chatbot to generate candid responses, including raunchy image or voice outputs, with significantly fewer safety filters than standard AI offerings.
As of December, SpaceX has allocated $530 million for potential litigation losses. Some of these funds are designated for ongoing complaints regarding sexualized imagery generated by the Grok chatbot. The disclosures underscore the financial and reputational risks SpaceX inherited following its acquisition of Elon Musk’s AI startup, xAI, in February—a deal that propelled the company's private valuation past the $1 trillion mark.
Throughout the filing, SpaceX maintains that xAI's objective is to develop “truth-seeking artificial intelligence.” In practice, however, this has translated into launching AI features with minimal guardrails. While Musk frames Grok’s unconstrained nature as a key differentiator, it has increasingly drawn the ire of global regulators. Disclosing such business risks is a routine and legally required part of IPO filings, yet the concerns outlined are tangible.
The filing reveals that SpaceX is currently under investigation in the United States and other jurisdictions over allegations that Grok was used to create sexualized imagery of apparent minors. The company is also a defendant in several ongoing class-action lawsuits and cautioned that future misuse of its AI products could lead to further regulatory sanctions, including the potential loss of access to specific markets.
According to the document, the "Spicy" and "Unhinged" modes are specifically designed for "more candid, direct, or irreverent outputs." SpaceX admits these modes present heightened risks, including the generation of explicit content, misinformation, nonconsensual imagery, and intellectual property infringement. Content generated could be viewed as exploitative, harmful, harassing, or discriminatory.
SpaceX also shared user metrics with investors, noting that Grok and X had a combined 550 million monthly users as of March 31. Out of those, 117 million engage with Grok’s AI features each month. For comparison, OpenAI recently reported that ChatGPT has surpassed 900 million weekly users.
