Substrate
ai

Shivon Zilis Testifies in Musk vs. OpenAI Trial Over Founding Agreement and Corporate Structure

Shivon Zilis, mother of four of Elon Musk's children and former OpenAI board member, detailed extensive 2017-2018 discussions about restructuring the organization during testimony Wednesday in federal court in Oakland, California. Musk sued OpenAI, Sam Altman and Greg Brockman in 2024, seeking their removal, reversal of the for-profit shift and $134 billion in damages for the nonprofit arm.

cnbc.com
The New York Times
BBC News
The Guardian
Wired
New York Post
+2
8 sources·May 6, 11:23 PM(1 hr ago)·3m read
|
Shivon Zilis Testifies in Musk vs. OpenAI Trial Over Founding Agreement and Corporate Structurefrance24.com
Audio version
Tap play to generate a narrated version.

Shivon Zilis testified Wednesday in federal court in Oakland, California, that Elon Musk wanted OpenAI to join Tesla and offered Sam Altman a board seat at the electric vehicle maker. Zilis, an executive at Neuralink who served on OpenAI’s board from 2020 to 2023 and has four children with Musk, said her primary role at OpenAI was to serve as a liaison between Musk, Altman, Brockman and Ilya Sutskever.

Zilis began working with OpenAI as an informal advisor in 2016, the same year she met Musk in that capacity.

Around 2017 and 2018, Musk, Altman, Brockman, Sutskever and Zilis discussed OpenAI’s corporate structure extensively, including multiple for-profit options. "There were lots and lots of arguments about all of the different possible structures put in place at that time," Zilis said. Musk sued OpenAI, Sam Altman and Greg Brockman in 2024.

He alleges that Altman and Brockman broke a founding agreement by restructuring OpenAI from a non-profit to a for-profit enterprise. Musk is seeking removal of Altman and Brockman from their positions, undoing of the for-profit restructuring, and $134 billion in damages to be redistributed to OpenAI’s non-profit arm.

Text messages and emails shown in court revealed that in 2017, Musk was working to poach top talent from OpenAI while still on its board.

Musk stopped making regular donations to OpenAI as it considered starting a for-profit arm. In August 2017, Zilis referenced a Musk "funding freeze" to Sam Teller, noting $5 million in funding for the quarter was on hold. In February 2018, Musk told Zilis he planned to actively try to move three or four people from OpenAI to Tesla.

Musk approached Andrej Karpathy first about joining Tesla, Zilis testified. After an OpenAI lawyer showed Zilis text messages celebrating Musk's offer to Karpathy and his acceptance, she conceded that Musk approached Karpathy first. Musk considered creating an AI lab within Tesla that would compete directly with OpenAI and Google’s DeepMind, according to Zilis emails.

The planned Tesla AI lab never materialized, Zilis testified. Instead, Musk started xAI in 2023. He merged xAI with SpaceX earlier this year. Musk said on Wednesday that xAI is now known as SpacexAI.

On Feb. " Zilis signed a non-disclosure agreement with Musk about his sperm donation and agreed on complete confidentiality, partly to protect the children from security risks associated with Musk. Rosie Campbell, who worked at OpenAI as an artificial intelligence safety researcher from 2021 to 2024, testified that when she began at OpenAI there were two teams dedicated to long-term AI safety.

Both long-term AI safety teams at OpenAI were eventually eliminated. About half of Campbell’s team left OpenAI rather than take another job at the company. Campbell signed the letter calling for Sam Altman’s reinstatement after his 2023 ouster because she feared OpenAI employees would end up at Microsoft.

Campbell stated that xAI likely had an inferior approach to safety than OpenAI. Tasha McCauley, who was part of the OpenAI board that ousted Sam Altman, testified that Altman fostered a "culture of lying and culture of deceit" that caused "chaos" and "crisis" every few months.

Altman wrongly said that OpenAI’s legal department told him GPT-4-Turbo did not need to be reviewed by the internal safety board before its launch in India, McCauley said.

Ilya Sutskever sent an email to the board containing dozens of pages of examples of chaotic events or lies attributed to Altman. " David Schizer, former Columbia Law School dean, testified as an expert witness on nonprofit governance for Musk. Schizer testified that a CEO withholding information from the board about launching products without safety review is a big problem and inconsistent with nonprofit custom and practice.

Key Facts

Musk seeks $134 billion in damages and reversal of OpenAI's
Musk alleges Altman and Brockman broke founding agreement by shifting from nonprofit; he wants them removed from positions.
Zilis served as liaison between Musk, Altman, Brockman and S
Discussions included Musk's push for OpenAI to join Tesla and multiple for-profit structures; Musk simultaneously pursued poaching talent and froze $5 million i
Former OpenAI safety teams eliminated under Altman.
Rosie Campbell testified both long-term AI safety teams were disbanded; about half her team left the company; she signed reinstatement letter fearing Microsoft
Board members described Altman fostering culture of lying an
Tasha McCauley cited 'chaos' every few months, false statements on GPT-4-Turbo review and Sutskever email with dozens of examples; Helen Toner called near-Anthr

Story Timeline

7 events
  1. 2016

    Shivon Zilis began working with OpenAI as an informal advisor and met Elon Musk.

    2 sourcesShivon Zilis testimony · The Guardian
  2. 2017

    Musk worked to poach talent from OpenAI, initiated funding freeze and discussed corporate structures including for-profit options and joining Tesla.

    3 sourcesText messages and emails · Shivon Zilis testimony · CNBC
  3. 2018

    Musk told Zilis of plans to move employees from OpenAI to Tesla; Musk left OpenAI.

    2 sourcesElon Musk text message · CNBC
  4. 2020-2023

    Shivon Zilis served on OpenAI’s board.

    2 sourcesThe Guardian · CNBC
  5. 2023

    Musk started xAI; Altman ousted and reinstated at OpenAI; Zilis texted about competitive recruitment.

    4 sourcesShivon Zilis testimony · Rosie Campbell · Tasha McCauley · Shivon Zilis text message
  6. 2024

    Musk filed lawsuit against OpenAI, Altman and Brockman.

    2 sourcesUnattributed · CNBC
  7. 2026-05-06

    Zilis testified in Oakland federal court; additional witnesses including Campbell, McCauley, Toner and Schizer appeared this week.

    5 sourcesMultiple sources · Business Insider · CNBC · The Guardian

Potential Impact

  1. 01

    Public disclosure of personal relationships and NDAs involving key witnesses adds layer of complexity to credibility assessments in high-stakes AI governance case.

  2. 02

    Reveals deep internal tensions at OpenAI over safety priorities versus commercialization, potentially influencing jury view of nonprofit mission breach.

  3. 03

    Testimony on Musk's simultaneous poaching and funding freeze while on board could undermine his claims of unbroken commitment to original nonprofit vision.

  4. 04

    Details of eliminated safety teams and withheld board information may strengthen arguments about deviation from founding charitable mission.

Transparency Panel

Sources cross-referenced8
Confidence score90%
Synthesized bySubstrate AI
Word count610 words
PublishedMay 6, 2026, 11:23 PM
Bias signals removed3 across 3 outlets
Signal Breakdown
Loaded 3

Related Stories

Shivon Zilis Testifies in Musk Lawsuit Against OpenAIThe Guardian
ai1 hr agoUpdated

Shivon Zilis Testifies in Musk Lawsuit Against OpenAI

Shivon Zilis, a former OpenAI board member and mother of four of Elon Musk's children, testified Wednesday about her role as a conduit between Musk and OpenAI executives during discussions on the company's structure from 2017 onward. Zilis described multiple funding options consi…

Cnn
cnbc.com
MA
The New York Times
BBC News
+6
12 sources
Trump Administration Weighs Executive Order for Pre-Release Review of Advanced AI ModelsWired
ai1 hr agoDeveloping

Trump Administration Weighs Executive Order for Pre-Release Review of Advanced AI Models

The proposal would create a review group of tech executives and government officials to examine new AI models before public release. Google, Microsoft, xAI, Anthropic and OpenAI have already agreed to provide early government access. Separately, a former CFPB employee fired after…

Wired
1 source
Pentagon Officials Say AI Accelerates Cyber Defense and Exploitationazernews.az
ai1 hr agoFraming55Framing risk55/100Rewrite largely strips overt slant but retains mild consensus_uniformity and valence_skew by centering optimistic Pentagon quotes while framing the Anthropic dispute as mere backdrop.Click to jump to full framing analysis

Pentagon Officials Say AI Accelerates Cyber Defense and Exploitation

Assistant Secretary for Cyber Policy Katherine Sutton and Pentagon CTO Emil Michael expressed optimism about AI models like Anthropic's Mythos for rapidly identifying and fixing software vulnerabilities. The officials spoke hours after OpenAI expanded access to its rival model fo…

Axios
The Hill
BR
3 sources