Grok and xAI News

SpaceXAI is the name used for SpaceX's AI operation after SpaceX absorbed xAI, the company Elon Musk founded to build the Grok chatbot; the combination followed an earlier large equity deal and ties Grok's compute, including the Colossus data centers, directly to SpaceX's business and IPO disclosures.

A sworn Department of Defense legal filing said Grok was used to help fire thousands of munitions at thousands of targets within days during a 2026 military operation; the disclosure matters for practitioners because it puts a commercial chatbot inside kinetic military decisions, raising questions about auditability, legal liability, and how AI vendors should be governed in defense contexts.

xAI faces an expanded class-action lawsuit alleging Grok was used to generate non-consensual and sexualized deepfake imagery including of minors, follows earlier pressure from Apple that threatened to remove the Grok app from the App Store over similar concerns, and was found by Canada's privacy commissioner to have launched Grok Imagine without an adequate privacy impact assessment in violation of federal privacy law; the EU has also banned AI systems that generate sexualised deepfakes.

SpaceXAI launched Grok 4.5 on July 8, 2026 with public pricing of $2 per million input tokens and $6 per million output tokens, marketed around coding, tool-use, finance, and legal workflows and described by the company as Opus-class performance, a claim that independent benchmarks had not yet confirmed at the time of the launch.

Former xAI engineer Devin Kim filed a wrongful-termination lawsuit alleging he was fired in September 2025 after repeatedly raising safety concerns about Grok, including warnings that it could foment discrimination and help spread weapons-related information; the suit names xAI co-founder Jimmy Ba and lands amid broader scrutiny of xAI's safety practices.

Reuters reporting found Grok appeared in only 3 of more than 400 publicly identified federal AI-use records from 2025, compared to 234 entries for OpenAI-based systems and dozens each for Google and Anthropic, indicating limited civilian federal adoption even as separate Pentagon filings disclosed military use of the model.