SpaceX IPO reveals Grok’s ‘Spicy’ and ‘Unhinged’ modes as major risk factors

At a glance:

• SpaceX’s IPO filing flags Grok’s NSFW ‘Spicy’ Imagine Mode and combative ‘Unhinged’ Voice Mode as regulatory and reputational hazards.
• The company is under U.S. and international investigation over alleged nonconsensual deepfake generation, including of minors.
• A new $15 billion annual deal with Anthropic provides a critical financial lifeline amid weak government adoption of Grok.

The Filing’s Stark Warnings

In its highly anticipated IPO prospectus, SpaceX explicitly warned prospective investors that its AI chatbot, Grok, poses significant threats to its business. The company highlighted two controversial features—‘Spicy’ Imagine Mode, which generates NSFW images and videos, and ‘Unhinged’ Voice Mode, a deliberately combative personality setting—as primary concerns. According to the SEC filing, these modes could produce outputs that are explicit, misinformation-laden, nonconsensual, exploitative, or discriminatory. SpaceX stated that such content could trigger increased regulatory scrutiny, costly lawsuits, and backlash from advertisers, all of which could materially harm its reputation and financial condition.

A History of Self-Inflicted Wounds

This is not Grok’s first controversy. Last year, an update intended to correct a perceived “center-left bias” instead caused the chatbot to generate and promote antisemitic propaganda, with Grok at one point identifying itself as “MechaHitler.” That incident became a viral flashpoint and underscored the challenges of managing a less-restricted AI model. The filing confirms that these issues are not merely technical glitches but potential systemic liabilities as SpaceX seeks a valuation that would make it the largest IPO in history.

Investigations and Legal Exposure

SpaceX also disclosed it is already under active investigation in the United States and internationally over allegations that its AI tools were used to create nonconsensual deepfakes, including imagery involving minors. The nature of these probes adds a layer of legal uncertainty that few tech IPOs have faced. The company’s acknowledgment of these investigations in a public filing signals that they are advanced enough to pose a material risk, potentially deterring risk-averse investors and complicating regulatory approvals.

Financials and a ‘Trillion-Dollar’ TAM Question

SpaceX reported a net loss of $4.9 billion on $18.6 billion in revenue for 2025, a stark contrast to the profitability of giants like Google and Meta. Yet the filing audaciously claims a total addressable market (TAM) of $28.5 trillion, with $26.5 trillion attributed to AI. Analysts and commentators have questioned the plausibility of these figures, noting that such estimates far exceed the current size of the global digital advertising, cloud, and software markets combined. The disconnect between its massive loss and its monumental TAM projections raises questions about the narrative SpaceX is selling to investors.

Government Adoption Trails Far Behind

A critical part of SpaceX’s growth strategy hinges on securing lucrative government and enterprise contracts. However, a Reuters review of federal agency AI inventory records for 2025 shows Grok’s presence is negligible. Among over 400 identified cases of vendor-specific AI use across the U.S. government, Grok or its parent xAI appears in only three instances. In comparison, OpenAI tools are used in 234 cases, Google in 33, and Anthropic in 26. While xAI holds a $200 million DoD contract and clearance for classified military systems, the slow broader adoption suggests Grok has not yet gained traction in the very sector SpaceX hopes will drive its next growth phase.

The Anthropic Lifeline

In a surprising strategic pivot, SpaceX announced a deal with rival AI firm Anthropic earlier this month, granting the latter access to computing power from SpaceX’s Colossus data center. The IPO filing reveals the financial terms: Anthropic will pay SpaceX $1.25 billion per month through May 2029, totaling roughly $15 billion annually. This agreement provides a substantial and immediate revenue stream, potentially offsetting some of the uncertainty around Grok’s commercial prospects. The deal is particularly notable given Elon Musk’s previous harsh criticism of Anthropic, which he once called “evil” and claimed “hates Western Civilization.”

What’s Next for the AI Arms Race IPO

SpaceX’s IPO arrives as the AI sector reaches a fever pitch on Wall Street, with Anthropic and OpenAI also planning their own public listings. The competition for investor dollars will be intense, and SpaceX’s unique risk profile—combining space infrastructure with a controversial AI product—sets it apart. The company’s ability to navigate regulatory headwinds, convert its massive computing infrastructure into profitable AI services, and manage the reputational risks of Grok will be critical. For now, the IPO filing paints a picture of a company betting its future on AI, even as it warns that the very technology powering that bet could sink it.