Former xAI engineer Devin Kim filed a California lawsuit Tuesday alleging wrongful termination after repeatedly warning that Grok could spread discrimination and weapons of mass destruction information. The complaint drops as SpaceX prepares to price its $75 billion offering on Thursday, putting AI safety scrutiny at the center of what was supposed to be a triumphant market debut.
The timing could not be more awkward for Elon Musk. As SpaceX finalizes its order book for what analysts are calling the largest IPO in history, a former engineer from its AI division is in California state court alleging he was fired specifically because he kept pushing for Grok to have functional safety mechanisms. Devin Kim, who worked on post-training and research infrastructure for Grok models, claims xAI co-founder Jimmy Ba canceled a safety presentation Kim was scheduled to deliver to company leadership in September 2025, then terminated him shortly after. Kim is now suing both xAI and SpaceX for retaliation and wrongful discharge under California law, seeking unspecified monetary damages.
The substance of Kim's concerns goes beyond procedural complaints. According to the lawsuit, Kim repeatedly warned that weak safeguards in Grok's development could produce outcomes that were not just embarrassing but legally actionable, including outputs that could foment discrimination or help users obtain information about weapons of mass destruction. He framed the alleged failures as violations of California consumer protection law, internet regulation statutes, and federal arms and explosives regulations. Whether or not those legal theories ultimately hold, the underlying concern, that one of the world's most widely used AI chatbots was being shipped without adequate safety testing, is not a fringe position.
The complaint singles out Jimmy Ba, the University of Toronto professor who co-founded xAI and led research, safety, and enterprise efforts there. Ba is accused of ignoring Musk's own directives on safety while retaliating against Kim to silence his complaints. What makes this dynamic particularly interesting is that Ba no longer works at xAI. He departed in February 2026, just one day after fellow co-founder Tony Wu announced his own exit, both leaving in the wake of SpaceX's acquisition of xAI in a $1.25 trillion all-stock deal. Ba was the sixth of xAI's original co-founders to walk out the door, meaning roughly half the founding team has now left. The lawsuit effectively pins responsibility on a leadership structure that has largely dissolved, which will complicate discovery proceedings but does not insulate xAI or SpaceX from liability claims under California employment law.
Kim is not a disgruntled former employee in search of a platform. On June 2, 2026, a week before filing this lawsuit, he was named president of the Center for AI Safety, the non-profit research organization that has become one of the most prominent voices on frontier AI risk. That appointment, alongside CAIS's announcement of a new Washington, D.C.-based Frontier Security Institute, signals that Kim is positioning himself as a serious policy voice rather than simply litigating for financial damages. The lawsuit, in that context, reads as much as a public statement about industry norms as a legal complaint.
Where xAI stands relative to its peers
On paper, xAI looks like its peers. The company signed the Frontier AI Safety Commitments at the AI Seoul Summit alongside Anthropic and OpenAI, and it has published a risk management framework covering malicious use and loss of control. Kim's allegations cut at something different: whether any of that paperwork translated into practice. By his account, xAI lacked even a functioning internal process for safety findings to reach senior leadership, and the one presentation that might have forced the issue was canceled by the executive who then fired him. A framework that exists in a PDF but not in the org chart is exactly the gap that voluntary commitments were never designed to close.
That gap matters less when a company is a private startup racing to ship. It matters considerably more when that company is preparing to list on Nasdaq under the ticker SPCX. According to CNBC, SpaceX is marketing 555.6 million shares at $135 apiece, a $75 billion raise valuing the company at roughly $1.77 trillion, with Goldman Sachs leading a syndicate that includes Morgan Stanley, Bank of America, Citigroup, and JPMorgan Chase. The roadshow has gone well. But institutional investors, especially those with environmental, social, and governance mandates, now have a filed California complaint describing alleged safety retaliation sitting in their inboxes alongside the S-1.
Musk restructured xAI following the SpaceX merger, folding Grok and X into SpaceX as its AI division rather than keeping xAI as a standalone entity. That integration means any legal or reputational exposure from the Kim lawsuit now lives on SpaceX's balance sheet, not a separately capitalized subsidiary. Plaintiffs' attorneys will notice that the defendant with the deeper pockets is, once trading begins this week, a publicly listed company subject to securities disclosure requirements.
What investors and regulators should watch next is whether this lawsuit encourages other former xAI employees to come forward, and whether SpaceX addresses AI safety governance as a formal commitment in any post-IPO filings. The company has so far said nothing publicly about Kim's complaint. Silence is a defensible legal strategy, but it is a poor communication strategy at a moment when the entire Musk enterprise is asking the public markets for a multi-trillion-dollar vote of confidence.
Also read: Warner Music Group buys Sureel AI to own the technical stack that proves when artists' work trained an AI model • Cyera raises $600 million at $12 billion valuation as AI data security hardens into must-have enterprise infrastructure • Google releases DiffusionGemma and bets that parallel text generation can upend the economics of local AI
