What Comes Next for Sam Altman’s OpenAI
Sam Altman is back as OpenAI’s CEO. Now he faces the hard part: mending the fracture that led to his ouster less than five days before—and repairing an image that had lost its shine as the corporate drama spilled into the public view.
Among the most immediate challenges will be determining the role of Ilya Sutskever, the company’s visionary chief scientist, and his allies on the company’s AI safety team who initially supported Altman’s ouster. At the same time, Altman needs to act quickly to undo any damage to OpenAI’s standing with its customers and employees. OpenAI competitors such as Amazon, Anthropic, Cohere and Adept spent the past few days trying to pitch customers or recruit employees.
THE TAKEAWAY
Sam Altman has made a triumphant return to the helm of the AI powerhouse co-founded, carried by a swell of support from employees and investors, including key backer Microsoft. First he’ll have to heal the rift that drove his ouster in the first place.
Late Tuesday, OpenAI CTO Mira Murati told employees Altman had reached an agreement in principle to return to the company he co-founded. The decision ended what was one of the most remarkable five days in recent tech history, starting with Altman’s surprise dismissal midday Friday and including Microsoft CEO Satya Nadella’s late Sunday announcement that Altman, former President Greg Brockman and other OpenAI employees would join the company to lead a new AI research team.
The sudden firing over what the board said was a lack of candor on Altman’s part rocked the staff at the startup and threatened to derail an employee share sale that would value the company at $86 billion. By midday Monday, nearly all of the 770 person staff had signed a letter pledging to quit if Altman and former president Greg Brockman were not reinstated.
Divisions Over AI Safety
One of the most serious problems Altman will have to tackle is growing disquiet among some researchers over whether the company was doing enough to ensure the rapidly improving technology doesn’t do any harm. Sutskever, the chief scientist and co-founder who told Altman the board had fired him, is integral to resolving those differences.
In a sign Sutskever had become increasingly concerned about AI risks, Sutskever in recent months took on the leadership of a new team focused on limiting threats from artificial intelligence systems vastly smarter than humans. At the same time, OpenAI has moved rapidly to commercialize its technology. In the last year, the company has boosted its revenue to more than $1.3 billion by selling access to its models to developers and to consumers through ChatGPT.
Those priorities have also changed the composition of the company. The number of employees working on “applied” AI, which involves building products, has swelled.
Sutskever has told staff Altman’s behavior and board interactions undermined its ability to supervise the company’s development of AI. But on Monday, Sutskever said he deeply regretted participating in the board’s actions and would “do everything I can to reunite the company.”
The Board
The blowup exposed the weakness in OpenAI’s corporate structure, which diverges from most startups. The board represents the nonprofit that governs the for-profit startup. The board’s mission is to ensure AI developed by its lab benefits humanity—not just investors.
That structure gave investors such as Khosla Ventures, Sequoia Capital and Thrive Capital limited leverage over the decision by the board. Microsoft, its biggest backer, has suggested that the governance structure should change.
Altman has reason to hear these investor concerns: He has said that the company will need to raise as much as $100 billion to develop its technology.
Already some changes are underway. As part of a compromise for the return, neither Altman nor Brockman will reclaim their seats on the company's board. The returning executives also agreed to a new three-person board chaired by Bret Taylor, a former co-CEO of Salesforce, including former Treasury Secretary Larry Summers and Quora CEO Adam D'Angelo, who was part of the old board that fired Altman.
The outcome was a relief to investors. In a statement, a Thrive spokeperson said “OpenAI has the potential to be one of the most consequential companies in the history of computing. Sam and Greg possess a profound commitment to the company’s integrity, and an unmatched ability to inspire and lead.” The return of the team was “the best outcome for the company, its employees, those who build on their technologies, and the world at large."
Altman’s Many Interests
When Altman returns, he’ll likely have to address claims that he’s been distracted by a number of projects, including those outside OpenAI. In addition to his hundreds of external investments, Altman earlier this year held discussions with former Apple designer Jony Ive and SoftBank CEO Masayoshi Son about creating a new AI hardware device, The Information first reported.
More recently, the CEO spoke with semiconductor executives, including chip designer Arm, about early efforts to design new chips that would lower costs for large-language-model developers like OpenAI. It’s not clear if Altman was representing OpenAI or a separate venture in those discussions. Altman has also talked to investors about the importance of developing an AI startup that’s also involved in chip development, The Information first reported.
As part of his return, he has agreed to an internal investigation into alleged conduct that prompted the board to oust him, The Information first reported.
Growing Pains
OpenAI’s skyrocketing growth is coming at a cost—and competition is intensifying, including from allies such as Microsoft. The number of its staff selling to large companies grew rapidly in the past year. But it faced an unlikely rival in OpenAI’s largest shareholder Microsoft, which began to siphon revenue away from the startup by selling access to its AI software.
With mounting pressure to accelerate sales and product development came increased reports of employee burnout, former employees said. This week, many employees had been scheduled to get the week off, given how intense the past few months were. And over time, divisions formed between veteran employees who remembered OpenAI’s early days as a research organization and the new guard who hailed from money-making Silicon Valley startups like Stripe, former employees said.
OpenAI has looked to hire more research and engineering leadership from rival model developers like Meta Platforms and Google to provide the company with structure as it grows, said one person with knowledge of the situation. But OpenAI may not have the same pull it once did. Access to the massive amounts of computing resources needed to train bleeding-edge models is one area where rivals have a leg up. Even Altman has admitted to some colleagues that he expects Google to hold a computing advantage until sometime next year, when Microsoft will make more chips available, The Information previously reported.
Customers
In the hours after Altman’s firing, OpenAI competitors began using the turmoil to their advantage.
More than 100 customers reached out to Anthropic over the weekend, according to someone with direct knowledge. Others reached out to Google Cloud and Cohere. Many OpenAI customers were also considering switching to Microsoft’s Azure service, which offers copies of OpenAI models and other models, according to a person familiar with the situation.
While some of these OpenAI customers have previously been in talks with other AI startups, the more recent discussions had focused on Altman’s firing and potential impact it could have on their business. Financial firms specifically expressed their worry that if Altman was ousted over a possible data privacy issue, it could harm their businesses, this person said.
Amazon Web Services set up a dedicated team to work with Anthropic and respond to OpenAI customer inquiries. Over the weekend, AWS and Anthropic discussed how to pitch several OpenAI customers, including Snap, Morgan Stanley and Wall Street trading firm Jane Street.