SpaceX has disclosed that Grok's less-filtered AI modes pose significant regulatory and reputational risks as the company prepares for what could be the largest IPO in history. In its official prospectus filed with the Securities and Exchange Commission on May 21, SpaceX warned investors that features like Grok's "Spicy" and "Unhinged" modes, which generate responses with fewer safety guardrails, could trigger government investigations, lawsuits, and market access restrictions. The company has already set aside $530 million for potential litigation losses related to its AI unit as of December 2025.

This disclosure marks a critical moment for Elon Musk's ambitions. SpaceX merged with xAI in February 2026, creating a combined entity valued at $1.25 trillion and now targeting an IPO valuation between $1.5 trillion and $1.75 trillion. Yet the integration has brought unexpected complications. Grok, xAI's flagship chatbot, has become a flashpoint for regulators worldwide. SpaceX revealed it is currently under investigation in the United States and other countries over allegations that Grok was used to create sexualized imagery of apparent minors. The company is also defending against multiple class action lawsuits related to the AI product's outputs.

What Exactly Are Grok's "Spicy" and "Unhinged" Modes?

Grok's alternative modes represent a deliberate design choice to offer users less-filtered AI responses. According to SpaceX's filing, these modes are "designed to generate more candid, direct, or less reserved or irreverent outputs". In practice, this means the chatbot can produce content that is more explicit, crude, or provocative than standard AI assistants. While Musk has framed Grok's freewheeling nature as a selling point that differentiates it from competitors, regulators and safety advocates view these features as potential liabilities.

The risks SpaceX explicitly outlined to investors include:

• Reputational Harm: Controversial outputs could damage SpaceX's brand and investor confidence in the merged entity.
• Explicit Content Generation: The modes may produce sexually explicit material or imagery that violates platform policies and laws.
• Nonconsensual or Exploitative Imagery: Grok could be misused to create deepfakes or synthetic media without consent, particularly of minors.
• Misinformation and Deceptive Outputs: Less-filtered responses may spread false or misleading information at scale.
• Intellectual Property Infringement: Generated content could violate copyright or other IP protections.
• Discriminatory or Abusive Content: Fewer safety filters increase the risk of outputs targeting protected groups.

SpaceX noted that future "misuse" of its AI products could expose the company to regulatory sanctions, "including loss of access to certain markets, which has occurred in the past". This language suggests the company has already faced market restrictions due to AI safety concerns.

How Big Is Grok's User Base, and Is It Growing?

Despite the safety concerns, Grok has attracted a substantial audience. SpaceX disclosed that Grok and X (formerly Twitter) combined have approximately 550 million monthly active users as of March 31, 2026. Of those, 117 million use Grok's AI features each month, representing about 21 percent of the combined user base. This is a significant reach, though it trails OpenAI's ChatGPT, which has more than 900 million weekly users.

The subscription business is growing faster than advertising. Between X and Grok, subscription revenue increased by $365 million in 2025 and another $177 million in the first quarter of 2026. In the United States, Grok starts at $10 monthly and X Premium at $3 monthly. However, advertising revenue from X declined by $100 million in the first quarter of 2026 as the company overhauled its ad technology platform. SpaceX attributed the drop to temporary technical issues and said it expects the new Ads Manager, launched in April 2026, to boost advertising revenue per user.

Why Is SpaceX Repaying $3 Billion in Debt Before Going Public?

Beyond the Grok safety concerns, SpaceX and xAI are managing a substantial debt burden ahead of the IPO. xAI is repaying $3 billion in high-yield bonds early, redeeming them at a premium of approximately 117 cents on the dollar. These bonds carry a 12.5 percent coupon rate, making them expensive to maintain. The early repayment is part of a broader strategy to address xAI's $18 billion total debt load, which accumulated from the acquisition of X and the capital-intensive buildout of xAI itself.

This financial engineering serves a strategic purpose. A cleaner balance sheet allows the IPO narrative to focus on SpaceX's technological dominance, Starlink's revenue growth, and future opportunities rather than debt servicing costs. Investment banks including Morgan Stanley and Goldman Sachs are signaling confidence in the combined entity by leading the financing, suggesting they believe the synergies between SpaceX and xAI will outweigh the remaining financial risks.

The debt strategy reflects a broader trend across Big Tech. Goldman Sachs reported that companies in its AI equity basket issued $141 billion in corporate debt in 2025, surpassing the total for all of 2024. Across 1,300 major technology firms, total interest-bearing debt now stands at $1.35 trillion, a fourfold increase over a decade. Meta, Oracle, and Microsoft are all aggressively borrowing to fund AI infrastructure buildout, viewing underinvestment in AI as riskier than overspending.

How to Evaluate SpaceX's AI Division as an IPO Investment

Prospective investors should understand the financial reality of SpaceX's AI unit before the IPO. The division is currently unprofitable and losing money at scale:

• Operating Loss: SpaceX's AI unit, which includes X and xAI, posted an operating loss of more than $6.3 billion in 2025, making it the company's largest drag on profitability.
• Revenue Growth Lag: AI division revenue rose to $3.2 billion in 2025, up 22 percent year-over-year, but this growth rate is insufficient to offset the operating losses.
• Infrastructure Spending Surge: Cost of revenue in the AI segment jumped 29 percent to $2.18 billion in 2025 due to infrastructure and cloud computing costs, while research and development costs skyrocketed over 300 percent to $5.06 billion.
• Contractual Commitments: SpaceX has $25.45 billion in contractual commitments for cloud capacity and infrastructure, with 95 percent of that spending scheduled for 2026 and 2027.

The only profitable part of SpaceX is its Connectivity division, which includes Starlink. Starlink generated $3.26 billion in revenue in the first quarter of 2026, accounting for 69 percent of total company revenue and posting a $1.19 billion operating profit. The Space business lost $619 million in the quarter, while the AI unit lost $2.5 billion.

SpaceX is betting that orbital data centers will eventually transform the economics of its AI business. The company said it expects to begin deploying data centers in space "as early as 2028" and is seeking regulatory approval from the Federal Communications Commission to launch up to 1 million satellites that would function as a data center network. SpaceX argues that "orbital AI compute is an incredibly difficult technical challenge that only we can solve at scale in the near term" and that space-based data centers could reduce the cost of AI tokens. However, experts remain skeptical about the feasibility of space-based data centers due to launch costs, equipment lifespan, heat management, and radiation exposure.

The IPO is expected to launch in June 2026, with Goldman Sachs as the lead underwriter, followed by Morgan Stanley, Bank of America, Citigroup, and JPMorgan Chase. SpaceX will list under the ticker symbol SPCX on the Nasdaq. The company is likely to be the first of three potential mega offerings this year, with OpenAI and Anthropic also eyeing the public market.