When Anthropic announced its $124 million Series A round on May 28, 2021, the investor roster read like a who's who of tech luminaries and philosophical philanthropists rather than the expected lineup of Sand Hill Road firms. Skype co-founder Jaan Tallinn led the round. Facebook co-founder Dustin Moskovitz participated. So did former Google CEO Eric Schmidt. Noticeably absent? Traditional venture capital.
For a raise that size, the composition was striking. Institutional investors typically dominate nine-figure Series A rounds, but Anthropic's early backers came almost entirely from circles aligned with effective altruism—a movement focused on using data-driven reasoning to maximize global good. The Center for Emerging Risk Research, an impact investor tied to that movement, joined Tallinn, Moskovitz, Schmidt, statistician James McClave and several undisclosed backers, according to the company's announcement that spring.
Secondary market data from Forge Global later pegged the post-money valuation at $623.11 million, or $2.57 per share, though Anthropic itself never confirmed those figures publicly.
The startup had been founded earlier that year by siblings Dario and Daniela Amodei, both former OpenAI executives. Dario took the CEO role; Daniela became president. Their pitch centered on what they called "steerable, interpretable, and robust" AI systems—academic language for making large language models safer and more controllable through techniques like reinforcement learning from human feedback.
"Anthropic's goal is to make the fundamental research advances that will let us build more capable, general, and reliable AI systems," Dario Amodei said in the May 28 statement. The phrasing was cautious, perhaps deliberately so given the team's experience watching AI development accelerate at OpenAI.
The Money and the Mission
The company chose to organize as a Public Benefit Corporation from day one, a Delaware legal structure that legally binds management to balance profit against a stated public mission. TechCrunch noted the PBC designation in its launch-day coverage. That structural choice signaled something about the founders' intent, though whether it would constrain or clarify decision-making remained an open question.
Moskovitz, whose wealth derived from Facebook and the productivity software company Asana, later told POLITICO that "any monetary returns from his investment in Anthropic will be entirely redirected back into our philanthropic work." Tallinn became a board observer. Schmidt, who had overseen Google's expansion into AI research during his tenure, joined as an individual backer rather than through a fund.
As Wired later reported, "most of the initial funding—$124M—came from sources affiliated with effective altruism," a framing that underscored just how much the round departed from Silicon Valley norms. Whether that insularity would prove an advantage or a limitation wasn't yet clear.
The funds were earmarked for computationally expensive research. Training large language models requires massive clusters of GPUs running for weeks or months, burning through electricity and cloud computing credits. Anthropic's technical agenda focused on interpretability—understanding what happens inside neural networks—and steerability, or guiding model outputs toward desired behaviors without unpredictable side effects.
What Came After
Anthropic returned for a $580 million Series B on April 29, 2022, pulling in several of the same backers: Tallinn, McClave, and the Center for Emerging Risk Research. That round also included Sam Bankman-Fried, the FTX cryptocurrency exchange, and its sister trading firm Alameda Research. Those names would acquire a different resonance after FTX's collapse later that year.
The startup's Series C in May 2023 marked a shift. Spark Capital led the $450 million round, bringing traditional venture firms into the cap table for the first time. More significantly, Google and Amazon made strategic commitments that would eventually total up to $6 billion across 2023 and 2024, though those deals came with conditions and complexity that went well beyond straightforward equity checks.
By the time Anthropic closed its Long-Term Benefit Trust structure in September 2023—granting the trust phased authority to elect and remove board directors through a special class of shares—the company had evolved far beyond its unusual Series A origins. Whether that early investor composition shaped the company's trajectory in meaningful ways, or simply reflected the Amodeis' personal networks at a moment when AI safety still lived on the margins of mainstream tech, remains a matter of interpretation.
What's undeniable is that $124 million bought Anthropic time to build without the pressure to chase conventional metrics. That breathing room, funded largely by philosophically motivated individuals rather than return-focused institutions, set a template the company would later abandon as it scaled into one of OpenAI's primary competitors.
