While OpenAI nears the finalization of a new funding round of approximately $100 billion, Anthropic has just secured a substantial financing of around $300 billion. More notably, at least 12 direct institutional investors in OpenAI have also emerged as backers in Anthropic’s latest fundraising effort.

The list of institutions placing bets on both companies includes not only hedge funds or asset management firms like D1, Fidelity, and TPG, which are accustomed to hedging their investments, but also, surprisingly, top-tier venture capital firms traditionally known for taking sides—such as Founders Fund, Iconiq, Insight Partners, and Sequoia Capital—have participated as well.

The most striking conflict of interest comes from global asset management giant BlackRock. Although BlackRock’s senior managing director and board member, Adebayo Ogunlesi, currently serves on OpenAI’s board, affiliated funds under BlackRock still took part in Anthropic’s $300 billion financing.

In public markets, it is common for asset management companies to hold shares in competing firms, and the broad hedging strategies of Microsoft and Nvidia across various AI companies are well-known. However, for venture capital, this represents a颠覆 of traditional norms.

VCs have long portrayed themselves as "founder-friendly" and "helpful." The core logic has been that when a VC holds a significant stake in a startup, it will fully support the company in competing against its main rivals. Additionally, as private entities, startups often share confidential business data with direct investors—information that public companies would not disclose.

Now, this boundary has become blurred. "If you own both OpenAI and Anthropic, to whom can you remain loyal, aside from your own limited partners?" one interviewed investor remarked bluntly, "As long as the firm doesn’t hold a board seat, people no longer see much harm in it."

OpenAI CEO Sam Altman, former president of Y Combinator, is well-versed in venture capital practices and had foreseen this trend. According to tech media outlet TechCrunch, Altman provided investors in 2024 with a list of OpenAI competitors he did not want them to support, including companies founded by former OpenAI employees, such as Anthropic, xAI, and Safe Superintelligence.

Although Altman later denied that supporting competitors would exclude investors from future OpenAI funding rounds, he drew a clear line when it comes to disclosing core interests. Documents revealed in Elon Musk’s lawsuit against OpenAI show that Altman explicitly informed investors: "If they made non-passive investments, they would no longer receive confidential business information from OpenAI."

The unique nature of the AI industry is breaking all established models. Major AI labs are experiencing unprecedented growth while facing record funding gaps for building data centers. Analysis suggests that when financing needs are so enormous and potential returns could be astronomical, it is difficult to expect investors to say "no."

Nevertheless, not all venture capital firms have slid down the slope of "double betting." Andreessen Horowitz (a16z) currently backs OpenAI and has not invested in Anthropic. Menlo Ventures supports Anthropic but has not invested in OpenAI. Firms like Bessemer, General Catalyst, and Greenoaks also appear to have directly invested in only one of the two companies for now.

However, it is undeniable that the breaking of this long-standing rule by respected Silicon Valley institutions like Sequoia Capital signals a fundamental shift in the market environment. For founders, regardless of where the term sheet comes from, conflict-of-interest policies must now be a key point of scrutiny before signing.