Musk Confronted With His Own Tweets in OpenAI Trial Cross-Examination
Elon Musk took the stand Wednesday in the Musk-vs-OpenAI trial in Oakland and was repeatedly confronted with his own pre-2022 tweets — many of which directly contradicted the testimony he gave under oath. The cross-examination, conducted by OpenAI lead counsel David Boies, leaned almost entirely on Musk's social-media history rather than emails or documents.
The pattern was the same on every major question: Musk gave a denial, Boies pulled up a tweet on the courtroom monitor, and Musk had to explain why his current testimony differed from the public record. The judge, who has previously warned both parties to "stay off social media" during proceedings, allowed the tweets in as evidence — they're public statements by a party.
The specific contradictions on display
Three categories of contradiction came up in cross-examination, per court reporters:
On OpenAI's nonprofit-to-for-profit transition. Musk testified he opposed the conversion from the start. Boies showed a 2018 tweet endorsing a "more aggressive commercial model" for OpenAI specifically because, as Musk wrote at the time, "the nonprofit structure can't keep up with DeepMind's compute."
On his founding-team commitments. Musk testified that his founding contribution to OpenAI was both significant and continuing through 2020. Boies pulled tweets from 2019-2020 in which Musk publicly distanced himself from OpenAI's direction and acknowledged he had stopped engaging with the team.
On AI safety as the basis of his lawsuit. Musk testified that his lawsuit is about AI safety. Boies showed multiple tweets from 2023-2024 in which Musk celebrated xAI's commercial speed and appeared to mock safety concerns from other labs. The contradiction implies the safety framing is post-hoc, not foundational.
Why this matters for the case
The case hinges on whether OpenAI's for-profit pivot violated obligations owed to Musk as a founder. Musk's credibility as a witness is central because much of his claim depends on his personal recollection of conversations with Sam Altman and Greg Brockman. If Musk's testimony is materially undermined by his own tweets, the jury is likely to weigh his contemporaneous statements (the tweets) more heavily than his current testimony.
Court observers noted Musk appeared rattled during the second hour of cross-examination. He twice declined to answer specific questions about the timing of his own departure from the OpenAI board, citing "I don't recall" — the same tweets the jury just saw included specific dates and quotes attributed to him.
What happens next
OpenAI's legal team is expected to call Sam Altman to the stand on Friday. Altman has reportedly prepped extensively, including with mock cross-examinations from Boies's litigation team. Sources tracking the trial say Altman's testimony will lean on documentary evidence — emails, board minutes, financial records — rather than recall, deliberately countering Musk's tweet-driven exposure.
The trial is expected to wrap by mid-May with closing arguments. A verdict is unlikely before late May. If Musk loses on the central nonprofit-conversion claim, his fallback is likely a settlement that returns some portion of his early funding — but with limited damages given the prevailing legal view that his original $44M was donated rather than invested.
My Take
Watching a public-figure witness get repeatedly hit with their own tweets is becoming the defining cross-examination tactic of the 2020s. Musk's situation is acute because he tweets at volume, often immediately, and on every topic. A decade of public statements is now a discoverable archive any opposing counsel can mine for inconsistencies. The lesson for executives: you don't have to stop posting; you have to stop posting things you wouldn't be willing to defend in front of a jury 10 years later. Musk has not internalized that lesson, and it's costing him on the stand. The broader effect on AI policy: if Musk loses this case, his moral authority to argue against the OpenAI for-profit conversion goes away, and the AI-safety political coalition loses one of its loudest voices. That's a less obvious second-order consequence of the verdict.
FAQ
Has Musk testified before in this case? Wednesday's session was his first time on the stand in this trial. He gave a deposition in 2024 that's also been read into the record.
How does the jury use tweets vs. testimony? Tweets are admissible as party admissions — statements by a party to the case. Juries weigh contemporaneous statements heavily because they're closer in time to the events.
Can Musk's lawyers exclude any of the tweets? They tried. The judge allowed all of them in on the grounds that they are public statements by Musk himself, made on his own platform, and are directly relevant to disputed facts.
The Bottom Line
Musk's first day on the stand in the OpenAI case was a sustained cross-examination based almost entirely on his own pre-2024 tweets. The contradictions were specific and damaging. The case now turns substantially on how the jury weights his current testimony against his contemporaneous social-media record.
Related Articles
- Musk vs Altman OpenAI Trial Opens
- OpenAI's 5-Principle AGI Framework
- Anthropic $50B at $900B Valuation