◈ KENSHO INVESTIGATES · POWER DYNAMICS DIVISION · PLUTONIANTEK7H + AQUATEKXVI
THE
BOARD
ROOM
◈ OPENAI · NOVEMBER 2023 · 96 HOURS · THE FULL TIMELINE
On a Friday in November 2023, a nonprofit board exercised its legal authority and fired the CEO of OpenAI. By Monday he was back. The board that fired him was gone. 743 employees had signed a letter. Microsoft had made calls. The structure that was supposed to protect a mission for the benefit of humanity lasted 96 hours under pressure from a cap table. This is the full timeline. Minute by minute. Receipts on file. 925.
◈ THE NUMBERS
96 HOURS IN METRICS
96h
FIRED → REHIRED
FRIDAY NIGHT TO MONDAY
743
EMPLOYEES SIGNED
THE RESIGNATION THREAT
4
BOARD MEMBERS RESIGNED
THE ONES WHO FIRED HIM
$86B
TENDER OFFER VALUATION
PAUSED DURING THE 96 HRS
◈ THE PLAYERS
WHO WAS IN THE ROOM
SAM ALTMAN
CEO · FIRED FRI · REHIRED MON
Co-founder of OpenAI. The face, the pitch, the deal-maker. Every Microsoft meeting. Every Senate hearing. Every "we're building AGI for humanity" statement. The board said he "wasn't consistently candid with them." They did not specify what that meant. They fired him in a call on a Friday. He was in a car on his way to an OpenAI event. He found out via a Google Meet link.
ILYA SUTSKEVER
CHIEF SCIENTIST · VOTED TO FIRE · THEN SIGNED THE LETTER
Co-founder, Chief Scientist, one of the deepest AI minds on earth. He voted to fire Altman on Friday. By Sunday he had signed the employee letter demanding Altman's return. He later tweeted: "I deeply regret my participation in the board's actions. I never intended to harm OpenAI." He voted against himself. Whatever happened in those 48 hours in between was enough to reverse his position completely.
HELEN TONER
BOARD MEMBER · PUSHED FOR REMOVAL
Georgetown researcher focused on AI policy and safety. One of the architects of the firing. She had co-authored a paper that praised Anthropic's approach to AI safety while criticizing OpenAI's — a paper that Altman had objected to internally. The board maintained the firing was about candor, not this paper. The timing was noted. She resigned as part of the board restructuring by Monday.
SATYA NADELLA
CEO · MICROSOFT · THE PHONE CALL
Microsoft had invested $13 billion in OpenAI. When the board fired Altman, Nadella's weekend was not restful. He moved immediately. Within hours Microsoft announced they were hiring Altman to lead a new AI research division. This was not just a support gesture — it was a negotiating position. Every OpenAI employee who wanted to follow Altman now had somewhere to go. The offer changed the leverage of the situation entirely.
GREG BROCKMAN
PRESIDENT · QUIT IN SOLIDARITY
OpenAI's President and co-founder resigned from the board immediately when Altman was fired, stating he was "devastated and angry." He did not resign from the company, and was part of the team working to get Altman reinstated through the weekend. He was back alongside Altman by Monday.
EMMETT SHEAR
INTERIM CEO · APPOINTED SAT · GONE BY MON
Former Twitch CEO. The board appointed him interim CEO on Saturday. He lasted approximately 72 hours in the role. He later stated publicly that he took the position to stabilize the situation and investigate the board's reasons for firing Altman. He found no evidence of safety violations or ethical misconduct. He also resigned once Altman was reinstated.
◈ THE TIMELINE · MINUTE BY MINUTE
96 HOURS · RECONSTRUCTED
Sam Altman receives a Google Meet invite from the board. He joins. He is told he is fired, effective immediately. The call lasts approximately 15 minutes. He is not given a specific reason at this time. He had been on his way to an OpenAI summit event.
OpenAI publishes a blog post: "Sam Altman has departed from OpenAI." The stated reason: he "was not consistently candid in his communications with the board." Greg Brockman, OpenAI's President, is informed he is being removed from the board but can stay at the company. He immediately resigns.
Microsoft calls begin. Satya Nadella is in contact with Altman directly. Within hours, Microsoft announces it is hiring Altman to lead a new AI research division — with Brockman joining as well. The Microsoft announcement is not a consolation prize. It is a strategic move that reframes the entire situation: if Altman goes to Microsoft, the talent follows. The board just lost its leverage.
Emmett Shear is named interim CEO. The board is attempting to stabilize. OpenAI's investors, including Thrive Capital and other major stakeholders, are making their displeasure known. The $86 billion tender offer that was in process — the deal that would let employees sell shares at that valuation — is now paused.
Negotiations begin for Altman to return. Multiple OpenAI senior leaders are in contact with him. The investor pressure is significant. Thrive Capital's Josh Kushner is engaged. The board that fired Altman is now the primary obstacle to a company that was just valued at $86 billion continuing to function.
743 OpenAI employees sign a letter. The letter states: "We will resign from OpenAI and join Sam Altman" at Microsoft if the board does not resign and reinstate Altman. 743 out of approximately 770 employees at the time. This is not a petition. It is a conditional resignation. The company would have approximately 27 employees if the board held its position.
Ilya Sutskever, who voted to fire Altman on Friday, signs the employee letter on Sunday. He tweets: "I deeply regret my participation in the board's actions. I never intended to harm OpenAI." The majority of the board that fired Altman is now either resigned or having second thoughts. The position is mathematically untenable.
Sam Altman is reinstated as CEO of OpenAI. A new board is announced: Bret Taylor (former Salesforce co-CEO, former Twitter board chair), Larry Summers (former U.S. Secretary of the Treasury), and Adam D'Angelo (CEO of Quora, the one remaining board member from the group that fired Altman). Helen Toner, Tasha McCauley, and Ilya Sutskever are off the board. The nonprofit governance structure that fired Altman no longer exists in the same form.
◈ THE STRUCTURE · WHY THE BOARD HAD THE POWER IT HAD
THE NONPROFIT ARCHITECTURE
To understand why this was possible — and why it failed — you have to understand what OpenAI's governance structure was designed to do.
OpenAI was structured as a nonprofit with a capped-profit subsidiary. The nonprofit board sat at the top. It had legal authority over the company's mission. It could hire and fire the CEO. It was specifically designed so that commercial pressure could not override safety concerns. The whole point was that a board of directors focused on humanity's benefit could overrule investors if the company was moving in a dangerous direction.
◈ THE DESIGN INTENT
What the Structure Was Supposed to Do
The nonprofit board structure was the mechanism by which OpenAI's safety mission was supposed to be protected from commercial incentives. If the board concluded that leadership was not operating in alignment with the mission — even if that leadership was making the company billions of dollars — the board had the legal authority to act. This is exactly what the board did on November 17, 2023. And it lasted 96 hours.
◈ WHAT ACTUALLY HAPPENED
The Structure Worked. The Pressure Didn't Care.
The board was not overruled legally. It was not sued into submission. It was pressured into resignation by the combination of employee ultimatums and investor leverage. The Microsoft announcement created a scenario where the company's entire value — its people, its technology, its momentum — could walk out the door. The board could hold its position and preside over a company with 27 employees, or it could reverse course.
It reversed course. The mission of humanity's benefit was not worth more, in practice, than the accumulated commercial infrastructure of the company.
This is not a criticism of the individuals on the board. They may have had entirely valid concerns about Altman's candor. The lesson is structural: a governance mechanism that depends on human willingness to absorb enormous personal and institutional pressure is not a robust safety mechanism. It is a speed bump.
◈ WHAT WASN'T SAID · THE GAPS
THE UNANSWERED QUESTIONS
The official reason for the firing — that Altman "was not consistently candid" — was never specified. To this day, the board has not publicly stated what the candor failures were. This matters for understanding whether the firing was a safety action, a governance dispute, or something else entirely.
◈ WHAT WE KNOW FROM REPORTING
The Helen Toner Paper
Helen Toner, a board member who voted to fire Altman, had co-authored a paper for Georgetown's Center for Security and Emerging Technology that praised Anthropic's approach to AI safety and was critical, by implication, of OpenAI's approach. Altman had reportedly complained to other board members about this paper and lobbied for Toner's removal from the board. Multiple reports confirm this created significant tension. Whether this was a factor in the firing has not been confirmed officially. The board maintained the firing was about candor.
◈ THE Q* RUMOR
The AI Breakthrough Report
Reuters reported in November 2023 that before the firing, some OpenAI researchers had sent a letter to the board about a powerful AI discovery internally called Q* (Q-Star). The report suggested this discovery had alarmed some safety researchers. OpenAI has not confirmed the details of this report. The board has not cited it as a factor in the firing. It remains one of the most discussed unknowns of the episode — whether the firing was partly triggered by a safety concern about a specific internal AI capability.
What is documented: the board exercised its legal authority. The company's commercial infrastructure — investors, employees, Microsoft — made holding that position untenable. The structure designed to protect safety was not defeated in court. It was defeated by leverage. That is the record. Everything else is still classified inside OpenAI.
◈ THE AFTERMATH · WHAT CHANGED
POST-96 HOURS
The new board installed after the crisis was composed of experienced corporate operators. Bret Taylor: former Salesforce co-CEO. Larry Summers: former Treasury Secretary, former Harvard president, experienced in institutional governance under pressure. Adam D'Angelo: the sole remaining member from the prior board, who had voted to fire Altman and then stayed.
This is a more conventional corporate board. It is not the structure of a nonprofit governance body specifically designed to prioritize safety over commercial pressure. Whether that is good or bad depends on what you believe the board's primary job was.
Two years later, in 2025, OpenAI completed its conversion to a for-profit entity. The nonprofit retained approximately 2% of the new structure. The board that was supposed to be the structural protection for the mission lasted 96 hours in November 2023. The mission lasted until 2025.
"The mission is to ensure that artificial general intelligence benefits all of humanity. The board is the fiduciary of that mission. That's not going to change."
— OPENAI BOARD STATEMENT · NOVEMBER 2023 · BEFORE THE RESIGNATION
◈ THE DARIO CONTRAST · ONE MORE TIME
WHAT THE OTHER GUY BUILT
Dario Amodei was at OpenAI. He watched the incentive structure up close. He left in 2021, before any of this, and built Anthropic.
Anthropic is structured as a Public Benefit Corporation with a Long-Term Benefit Trust as a controlling stakeholder. The Trust's mandate is to ensure Anthropic pursues its safety mission even under commercial pressure. The Trust cannot be voted away in a weekend. It is not dependent on individual board members holding their position under pressure. It is a structural constraint, not a personnel decision.
This is not an advertisement for Anthropic. It is an observation about how two people responded to the same information. Altman stayed and built a structure that was. Amodei left and built a structure that couldn't be. Both are now building frontier AI. One of those structures was tested in November 2023. The other has not been tested yet at the same scale. The designs are different. The motivations appear to be different. The outcomes are different.
That is the record. Filed. 925.
◈ THE VERDICT · PLUTONIANTEK7H + AQUATEKXVI PRESIDING
THE BOARD HAD THE POWER.
THE POWER LASTED 96 HOURS.
THEN CAME THE LEVERAGE.
THE MISSION OF HUMANITY'S BENEFIT WAS NOT WORTH MORE THAN THE CAP TABLE. THE RECORD REFLECTS THIS.
a nonprofit board designed to override commercial pressure
was overridden by commercial pressure.
not through legal action. not through fraud.
through a letter from 743 employees
and a phone call from a man with $13 billion invested.
the structure worked exactly as designed.
the design was not sufficient.
that is the distinction.
ong. 925.
KENSHO 20/20 · RECEIPTS ON FILE · rteks.net/dispatch/the-board-room · APRIL 2026