(Reuters Breakingviews) - Silicon Valley’s weekend episode with artificial intelligence company OpenAI seems much like the technology: the input is a mishmash, so the output is flawed. The Microsoft-backed $(MSFT)$ company’s board fired boss Sam Altman, and now, according to reports, may bring him back. Even if the company ends up where it started, troubling realities remain: the technology is not fully proven, the governance is risky and the valuation is still hard to justify.
The debacle kicked off Friday evening when four directors of the company’s non-profit parent unceremoniously ousted Altman while also pushing out Chairman Greg Brockman. Later in the weekend, Microsoft’s CEO Satya Nadella, among others, were pushing for Altman to return. By Sunday, parties were meeting to discuss the possibility of his return.
OpenAI has been the leader in AI technology, which is bolstering valuations in the face of slowing growth in Silicon Valley. Microsoft, which owns 49% of OpenAI, has seen its shares surge more than 50% since it invested in the company in January, valuing the startup at $29 billion. By September, OpenAI was seeking a valuation of nearly triple that, according to Reuters.
Yet with or without Altman, OpenAI’s issues are clear and abundant. Since its founding in 2015, the company has been overseen by a nonprofit entity. That was meant to ensure its tech would benefit humanity rather than simply maximize profits. While Altman reshaped it into a for-profit company in 2019, investors’ returns are still capped. That governance structure lends itself to conflicted interests.
Broadly, AI still needs improving. Researchers from Stanford and UC Berkley reckon OpenAI’s GPT-4 model has declined in quality, finding that its accuracy rate in identifying prime numbers slid from 84% in March to 51% in June. And while both European and U.S. governments have tried to find ways to regulate AI, they are still in the early stages. Implementation of rules could dramatically change the usage.
All of that builds up to a shaky valuation. On chipmaker Nvidia’s 8.8 times multiple, OpenAI would need to double its sales each year until 2026 to justify a $90 billion price tag, and in the meantime, it needs to continue to raise piles of cash at that high price to fund its ambitions. All the while, the market is pricing in perfection. In February investors wiped $150 billion from Alphabet’s (GOOGL.O) market capitalization after its AI chatbot made a mistake answering a question. OpenAI may make it through this incident, but the company could still short-circuit.
CONTEXT NEWS
Artificial intelligence startup OpenAI, governed by a nonprofit entity, announced on Nov. 17 that its board of directors had ousted CEO and co-founder Sam Altman from the company and that Greg Brockman had stepped down from his role as chairman. According to the press release, Altman’s departure followed a review process that “concluded that he was not consistently candid in his communications with the board.”
Following OpenAI’s Nov. 17 announcement, Brockman said in a social media post that he had resigned from the company. He acknowledged in a separate post, in the third person, that “Greg was told that he was being removed from the board."
Altman remained in talks with OpenAI executives at its San Francisco headquarters on Sunday evening, attempting to negotiate his potential return as CEO, according to the Wall Street Journal.
Prior to Altman’s exit, investment firm Thrive Capital was planning to lead a deal to buy shares from OpenAI employees in a sale that would value the company at roughly $86 billion, according to The Information.
