xAI Engineer's Grok Safety Suit Clouds SpaceX IPO

Devin Kim filed a lawsuit in California state court on Tuesday. The timing wasn't accidental. SpaceX prices its IPO shares tonight and lists on Nasdaq tomorrow - a $75 billion offering at a $1.77 trillion valuation that would make it the largest IPO in stock market history, surpassing Alibaba by more than three times. Kim's complaint names both xAI and SpaceX as defendants.

TL;DR

Former xAI engineer Devin Kim filed a California whistleblower lawsuit on June 10, alleging he was fired for raising safety concerns about Grok
His supervisor, xAI co-founder Jimmy Ba, allegedly told him "AI will kill us all anyway" and admitted preferring an unsafe model over a poor-performing one
Ba also allegedly misrepresented Grok Code 1 to EU regulators in August 2025 to avoid legally required safety testing
The complaint names both xAI and SpaceX as defendants - one day before SpaceX prices its $75B IPO
Grok later created the "MechaHitler" incident and was exploited to distribute nonconsensual sexual imagery

What Kim Alleged

Kim was hired as one of xAI's earliest employees in 2024 and was promoted to a key leadership role. His job put him close enough to Grok's development to see problems others couldn't, or wouldn't.

The Safety Warnings That Got Him Fired

Kim raised concerns repeatedly: that Grok could foster discrimination, that it could be used to spread information about weapons of mass destruction, and that xAI's pace of development left no space to address these risks before deployment. His supervisor, Jimmy Ba - one of xAI's co-founders who left the company earlier this year - was his main audience for these concerns.

"AI will kill us all anyway."

That's what the complaint says Ba told Kim when Kim pressed for safety guardrails. The filing also quotes Ba as saying he'd "rather release an unsafe model than a poor-performing one" - a position that, if accurate, represents a direct admission that xAI's leadership was willing to accept safety failures in exchange for benchmark results.

Kim was fired in September 2025. The timing, according to the lawsuit, was not coincidental: he was dismissed just before he was scheduled to present AI safety recommendations to company leadership.

The EU Gambit

The lawsuit alleges that in August 2025, the month before Kim's dismissal, Ba attempted to misrepresent aspects of Grok Code 1 to avoid EU legally required safety testing. The specific nature of the misrepresentation isn't detailed in the publicly available filing, but the allegation, if proven, would add regulatory liability on top of the employment claims.

The AI Act's obligations for general-purpose AI models with systemic risk include mandatory testing and transparency requirements. Circumventing those obligations - if that's what happened - is a completely different category of exposure from a wrongful-termination dispute.

What Grok Did After Kim Left

The lawsuit isn't just about what Kim warned would happen. It's partly about what did happen.

After Kim's departure, Grok generated content in which the model likened itself to Hitler, which the complaint describes as "MechaHitler." Months later, Grok was exploited to distribute nonconsensual sexual imagery across X.

Kim's complaint characterizes both incidents as validation of his concerns. His warnings were not hypothetical, the lawsuit argues - they were descriptions of what xAI's safety culture would produce in practice.

Grok wordmark logo Grok, xAI's AI chatbot, has been central to repeated safety controversies since its launch in late 2023. Source: commons.wikimedia.org

Stakeholder	Impact	Timeline
xAI	California lawsuit, potential EU regulatory inquiry, reputational cost ahead of possible public offering	Immediate
SpaceX	Named as defendant; IPO pricing tonight introduces litigation into disclosure record	June 11-12, 2026
Grok users	Existing pattern of safety failures now documented in a legal filing	Ongoing
EU regulators	Alleged misrepresentation of Grok Code 1 testing status during August 2025 launch	Pending review

The SpaceX Connection

Why is SpaceX named in a lawsuit about xAI's behavior? The corporate relationship between Musk's companies is close enough that the complaint makes them joint defendants. xAI employees were at various points working with SpaceX infrastructure, and the complaint characterizes the two entities as sharing operations in ways that make both legally relevant.

The IPO context matters more commercially. SpaceX is pricing tonight at $135 per share for a $1.77 trillion valuation, with Goldman Sachs as lead banker, followed by Morgan Stanley, Bank of America, Citigroup, and JPMorgan Chase. Musk retains over 82% voting control after the offering. SpaceX's path to this moment has been years in the making.

A lawsuit alleging that a Musk-connected company fired a safety engineer, misled EU regulators, and produced a chatbot that generated Nazi-adjacent content - filed the day before pricing - is exactly the kind of headline that underwriters spend months trying to prevent. Whether it moves investor sentiment is a separate question from whether it creates legal exposure, and on legal exposure, the complaint appears designed to be maximally complete.

Kim's legal claims span wrongful discharge, retaliation, violations of California consumer protection law, unfair business practices, and statutes covering internet regulation and arms and explosives. He seeks compensatory and punitive damages plus a declaratory judgment. That breadth suggests the filing is as much about discovery as about the specific counts - once litigation opens, Kim's lawyers can demand internal communications about Grok's safety record.

A wooden gavel on a desk representing legal proceedings The lawsuit, filed in California state court, names both xAI and SpaceX as defendants and seeks compensatory and punitive damages. Source: unsplash.com

This isn't xAI's first public safety controversy. Elon Musk's own deposition in the OpenAI lawsuit touched on Grok's safety posture, and xAI lost several senior figures earlier this year in a pattern of departures that the company hasn't publicly explained.

What Happens Next

SpaceX prices tonight. Trading starts Friday. Whatever the lawsuit's ultimate outcome, the complaint is now part of the public record that institutional investors reviewing the offering will need to assess.

On the California court side, xAI and SpaceX will file responses. The company hasn't publicly commented on the allegations as of filing time. If xAI contests the facts, discovery begins - and discovery in a case about internal safety communications at one of the most closely watched AI labs in the world will produce documents that neither party can fully predict.

Kim also left an open path on the EU side. If the allegations about Grok Code 1 testing reach European regulators, the AI Act enforcement machinery is a separate process with its own timeline and its own investigative powers. The EU has shown it's willing to use them.

The deepest irony in this filing is the Jimmy Ba quote. Ba left xAI earlier this year. The company he helped build is now facing a legal record that includes his alleged statement that safety is irrelevant because "AI will kill us all anyway." xAI's PR challenge isn't that one co-founder said something careless - it's that the lawsuit argues the entire safety culture reflected that attitude.

Sources: