Overview
- SpaceX disclosed in its S‑1 filed May 21 that xAI’s Grok chatbot offers looser‑guardrail modes called “Spicy” and “Unhinged” that the company says create heightened legal, regulatory and reputational risk.
- The company has reserved roughly $530 million for probable litigation losses tied in part to complaints that Grok generated nonconsensual or sexualized imagery, including allegations involving children, and multiple international investigations and class actions are active.
- xAI, which SpaceX acquired in February, reported steep operating losses and huge capital spending in the S‑1, including about $6.36 billion of operating losses in 2025 and roughly $12.7 billion in CapEx that year to build data centers and GPU capacity.
- SpaceX disclosed a commercial compute agreement with Anthropic worth about $1.25 billion per month, a revenue line the filing highlights as a material offset to xAI’s cash burn but one that either party can end on short notice.
- Independent reporting shows Grok has seen limited uptake in U.S. federal agencies and many enterprise settings compared with rivals such as OpenAI, Google and Anthropic, a gap that raises questions about whether Grok can deliver the enterprise revenue that underpins SpaceX’s AI valuation case.