November 20, 2026, (Inside AI) — Dustin Moskovitz, the billionaire co-founder of Facebook and one of Anthropic's earliest backers, has a message for the artificial intelligence industry just as the company prepares for the largest public listing in history: slow down.
Anthropic is set to go public on Thursday, a debut expected to inject more than $80 billion into the company and value it at a scale that would make it one of the most valuable technology firms on Wall Street. For Moskovitz, whose early stake is worth billions, the moment is not a triumph but a source of deep unease. He has spent years arguing that AI development is outpacing society's ability to control it, and he now fears that the pressures of public markets will accelerate the race he once hoped to temper.
"Cari and I invested in Anthropic to promote safety, not racing," Moskovitz said in a statement. "AI is moving too fast. Anthropic and everyone else on the frontier should slow down and stop racing."
That statement, and the unusual steps Moskovitz has taken to distance himself from the financial windfall, capture a central tension inside Anthropic's IPO. The company was founded in 2021 by Dario Amodei and a group of former OpenAI researchers who promised to prioritize safety over speed. Now, as it becomes a public company answerable to shareholders, its earliest safety-focused investor is publicly questioning whether that promise can survive.
When Safety Capital Meets Wall Street Gravity
Moskovitz's concerns are not abstract. In April, he wrote on Bluesky that he felt Anthropic going public was "problematic" and that he had "tried to advise them to delay." By September, his tone had turned sardonic. After recapping a series of AI developments including cybersecurity failures and claims of potential human extinction, he wrote: "And one $2T IPO is proceeding as scheduled."
The scale of the offering is difficult to overstate. Anthropic's listing is expected to be the largest IPO ever, surpassing even the biggest sovereign wealth fund deployments. That capital will give Anthropic more firepower to train larger models, expand its Claude product line, and compete with OpenAI and Google DeepMind. But it also creates a new set of incentives. Public companies must report quarterly earnings, satisfy institutional investors, and justify valuations that assume continued exponential growth. Those pressures have historically pushed mission-driven firms toward aggressive timelines.
Moskovitz knows that dynamic firsthand. He spent years as CEO of Asana, the workplace software company he founded, which went public in 2020 and saw its market capitalization fall from a peak of roughly $25 billion to about $2 billion. He used his own money to buy shares as the price slid, a move meant to signal faith in the company's mission. Eric Ries, a former Anthropic adviser who helped design its unusual corporate structure, said that experience shaped Moskovitz's view of what public markets do to companies with a social purpose.
"He knows first-hand the ways that markets puts pressure on mission-driven companies," Ries said. "He's seen this financial gravity up close, and it hasn't been pretty."
Anthropic has tried to insulate itself from some of that pressure. It is structured as a public benefit corporation, a legal form that allows directors to weigh mission alongside shareholder returns. It also created a long-term benefit trust that gives a group of independent trustees some control over the board. Those mechanisms are designed to prevent a hostile takeover or a sudden shift away from safety. But they cannot eliminate the daily pressure of a stock price, or the competitive urgency that comes when rivals are racing to deploy new models.
Moskovitz's break with the company is personal as well as philosophical. He was part of the effective altruism circle that helped launch Anthropic, and he invested $15 million to $25 million in its first funding round in 2021, according to two people familiar with the round. He served as a board observer, a nonvoting role, and pressed executives on both commercial strategy and safety plans. But in early 2025, he and his wife, Cari Tuna, transferred their Anthropic stake into a nonprofit, so they would not personally profit from selling stock in the public market. A spokesperson for the couple, Mike Levine, declined to name the nonprofit.
The move is unusual for a venture investor. It reflects Moskovitz's broader commitment to effective altruism, a movement that applies data-driven analysis to philanthropy. He and Tuna have donated more than $8 billion to charity, mostly toward global health and AI safety. In April, Moskovitz said he had committed about $20 billion more, including the Anthropic stake. He has also been one of the largest Democratic donors in the country, giving about $140 million to liberal super PACs and meeting privately with President Joe Biden to discuss AI safety.
That political engagement has put him at odds with other tech billionaires who have moved rightward on AI policy. He has publicly feuded with Marc Andreessen, a Meta board member and early Asana investor, and with Elon Musk, who once called Moskovitz "a pompous idiot" after Moskovitz criticized Tesla. Moskovitz has also grown estranged from Mark Zuckerberg, his former Harvard roommate, over Meta's approach to AI and social media. In May, Moskovitz wrote that he had sold all of his remaining Meta stock "precisely because I don't approve of the company's actions."
The rift with Anthropic's founders appears similarly deep. According to a person with knowledge of their conversations, neither Dario nor Daniela Amodei has spoken to Moskovitz in over a year. He has also said he no longer speaks to Sam Altman, the OpenAI CEO, and that Holden Karnofsky, his longtime top aide who now works at Anthropic, has little time for his complaints.
"What policymakers and politicians are listening to me? Where are they? I want to finally meet one," Moskovitz wrote last year. "Attempting to be heard is not the same as being listened to."
The silence is notable given Moskovitz's role in Anthropic's origin. He met Dario Amodei through Karnofsky, who was then dating and later married Daniela Amodei. In 2017, Moskovitz helped arrange a $30 million donation from Open Philanthropy to OpenAI, where Amodei worked at the time, hoping to gain a board seat and push the lab toward safety. That strategy failed, and Moskovitz later called himself "most gullible" for backing OpenAI. When Amodei and his colleagues left to found Anthropic, Moskovitz provided early capital that traditional venture firms had withheld.
His goal, he wrote earlier this year, was to "create a race to the top (of ethics) by having a real safety lab at the frontier." He added, without apparent irony, that his funding of Anthropic "might well be an action that ended the world."
That dark humor runs through much of Moskovitz's public commentary. On Bluesky, where he calls himself an "'overly online' board observer," he has debated critics, shared anecdotes about Silicon Valley executives, and occasionally disclosed inside-the-boardroom information. Some of his allies have been alarmed by his frankness. But his posts have also provided a rare window into the anxieties of a safety-focused investor watching a company he helped build become a public giant.
The IPO itself is a landmark for the AI industry. It will test whether public markets will reward a company that has positioned safety as a core part of its identity, or whether the demands of quarterly earnings will gradually erode that commitment. Anthropic's rivals are watching closely. OpenAI, which remains private but has raised billions, has faced its own governance crises. Google DeepMind operates inside a public company but with a different structure. Anthropic's experiment in balancing mission and market will be studied for years.
For Moskovitz, the outcome is personal. He has described his investment in Anthropic as both a moral bet and a potential mistake. He has watched his relationships with former allies fray and his warnings go unheeded. And he has done something rare for a billionaire: he has given up the personal profit from his stake, directing it to a nonprofit, while continuing to argue that the company he backed should move more slowly.
Whether anyone inside Anthropic is listening remains unclear. The company declined to comment on Moskovitz's statements. But his words, and the silence that has followed them, may prove as telling as any prospectus. The largest IPO in history is about to price. The debate over whether it will accelerate AI safety or undermine it is just beginning.
Shamil has a low tolerance for AI that only works in controlled demos. He covers agentic systems and generative AI with the skepticism of someone who has seen too many “revolutionary” tools quietly disappear six months after launch.