The Investors Who Bought Anthropic Before the World Understood Its Value

Anthropic has had at least two generations of unusually early investors.The first backed the company in 2021, before Claude existed, when Anthropic was a new research lab founded by former OpenAI researchers and its commercial prospects were largely theoretical. The second arrived three years later under very different circumstances. By 2024, Claude was already in the market, Amazon and Google had committed billions to Anthropic, and generative AI had become one of the defining investment themes in technology. Yet even then, Anthropic may have been dramatically undervalued. That second opportunity came partly because of an event that had nothing to do with Anthropic itself: the collapse of FTX. The bankrupt crypto exchange's estate needed to sell its stake in Anthropic, putting shares in one of the world's most closely watched private AI companies into the secondary market. Investors who bought during those sales were effectively purchasing Anthropic at a valuation of roughly $18 billion. Two years later, Anthropic raised $65 billion at a $965 billion post-money valuation. That rise helps explain why the company's investor history is becoming almost as interesting as its technology. Some investors identified Anthropic before there was a product. Others bought after the company was already prominent but before its valuation increased by more than an order of magnitude. The distinction matters because the two groups were making very different bets. Before Claude, a $124 Million Bet on Anthropic Anthropic's $124 million Series A in 2021 was led by Skype cofounder Jaan Tallinn and included Dustin Moskovitz, James McClave and former Google CEO Eric Schmidt. At roughly a $623 million post-money valuation, the company was still closer to a research venture than the commercial AI company it would later become. There was no Claude chatbot for investors to test. There were no multibillion-dollar cloud partnerships. ChatGPT had not yet triggered the generative AI boom. The investment case rested largely on the founders, the research and a belief that increasingly powerful AI systems would require a different approach to safety and reliability. Tallinn returned in Anthropic's 2022 Series B. Moskovitz later transferred his Anthropic position into a nonprofit vehicle, meaning any eventual gains from the investment could support philanthropic work. McClave stands out for a different reason. Public venture databases show just two disclosed investments associated with him, both in Anthropic. That does not establish the entirety of his portfolio, but it does show how prominently Anthropic appears in his publicly recorded venture activity. Former Google CEO Eric Schmidt was another participant in the early round, adding one of the technology industry's most recognizable names to Anthropic's initial investor base. Anthropic also attracted people who understood the technical requirements of building frontier AI. DeepMind cofounder Demis Hassabis was later reported to have been an early angel investor. Anjney Midha, another early Anthropic backer, focused on a constraint that would become central to the AI industry: access to computing power. Before the enormous infrastructure spending now associated with frontier models became obvious to the broader market, Anthropic needed both capital and access to the compute required to train them. Those early investors were underwriting a company whose eventual scale was still hypothetical. By 2024, the question had changed. FTX Created a Very Different Opportunity Anthropic's 2022 Series B included investments connected to Sam Bankman-Fried, Caroline Ellison and Nishad Singh. Then FTX collapsed. Its bankruptcy estate inherited a valuable asset: a substantial Anthropic stake acquired before the explosion in demand for generative AI. The estate ultimately sold its Anthropic position for roughly $1.3 billion, compared with an original investment of about $500 million. From the perspective of the bankruptcy estate, that was a significant gain. What happened afterward makes the transaction more striking. The sales took place when Anthropic was valued at roughly $18 billion. At that point, the company was no longer obscure. Claude had launched. Amazon and Google had established major relationships with Anthropic. Investors no longer needed to believe that generative AI might become an important technology category. They had to decide whether an $18 billion valuation still underestimated what Anthropic could become. Several investors did. Craig Falls Bought About $20 Million Personally One of the clearest examples is Craig Falls, the former head of quantitative research at Jane Street. Falls purchased 666,667 Anthropic shares for approximately $20 million during the FTX estate sale. The transaction provides something most private-company investment stories do not: a disclosed purchase price and share count. That makes the comparison unusually concrete. An $18 billion company already sounds large by conventional venture-capital standards. The question facing Falls was not whether Anthropic was an undiscovered startup. It was whether a company already valued in the tens of billions could still be priced far below its eventual scale. Anthropic's subsequent financing suggests why that distinction matters. Jesse Leimgruber Bought Even Though Anthropic Looked Expensive Entrepreneur and investor Jesse Leimgruber also acquired Anthropic shares through the FTX bankruptcy process. He later said the investment had appreciated about 5,400% in two years. Perhaps the most interesting part of his account is that Anthropic did not necessarily look cheap when he bought it. That is an important feature of the company's secondary-market story. Investors were not buying a forgotten asset at a distressed startup valuation. They were purchasing shares in a prominent AI company whose valuation had already risen substantially. The opportunity existed because their view of Anthropic's future value differed sharply from the price available at the time. Jason Haider Brought a Different Background to the Trade Jason Haider, founder and CEO of medical technology company Xenco Medical, also appears among investors associated with Anthropic during this period. His involvement is notable because he came from outside the relatively small group of investors and operators already embedded in frontier AI. Public disclosures do not establish enough about his position to calculate his return or determine precisely how much Anthropic exposure he still holds. That limitation applies to several investors connected to the company and is important when evaluating the fortunes that may ultimately emerge from Anthropic's rise. What can be established is timing. Haider was among the private investors gaining exposure during a period when Anthropic shares were available at valuations far below the company's later financing price. A Broader Group Entered the Anthropic Cap Table Trevor Paladino, Sahin Boydas and Pamela Hsieh also appear among individual investors associated with Anthropic as the company expanded. The available public record does not show the same transaction detail for every investor, so their positions should not be treated as equivalent to purchases where price and share count are known. Together, however, they illustrate a change in who was able to participate in Anthropic's growth. The company's earliest financing came largely from investors with direct ties to technology, AI research and venture capital. As Anthropic matured, secondary transactions created opportunities for a broader group of private investors to acquire exposure before the company reached public markets. Jared Carmel and Manhattan Venture Partners represent the institutional side of that market. The firm's participation in Anthropic fits a strategy centered on buying shares of large private technology companies before an eventual public offering. Anthropic is an unusually powerful example of why that market exists. At an $18 billion valuation, the company was already far beyond what most investors would call an early-stage startup. Relative to a later $965 billion financing, however, that same valuation looks dramatically different. Anthropic's Cap Table Tells Two Stories The easiest way to tell Anthropic's investment story is to focus exclusively on the investors who arrived first. They took the greatest technological risk. They invested before Claude, before the commercial proof and before generative AI became a global investment theme. But Anthropic's later investors may offer the more unusual lesson. By 2024, the basic secret was gone. The world knew frontier AI mattered. It knew Anthropic was one of the few companies capable of competing near the front of the field. It knew Amazon and Google were willing to commit billions. What remained uncertain was scale. An $18 billion valuation required investors to believe that a company already considered expensive could still be profoundly mispriced. For at least some of the people who bought Anthropic shares during that period, that appears to have been the more consequential insight. The final fortunes will depend on factors that remain unknown, including how many shares individual investors still own, subsequent dilution, additional sales and the price Anthropic ultimately commands if it enters the public market. That means it is too early to declare exactly who made what. But the documented transactions already tell an extraordinary story. Some investors bought Anthropic before Claude existed. Others bought after almost everyone knew the company mattered. Both groups may turn out to have been early.