During his appearance at the Morgan Stanley Technology, Media and Telecom conference in San Francisco on Wednesday, Nvidia CEO Jensen Huang indicated that his company’s recent investments in OpenAI and Anthropic are likely to be its last in both firms. He noted that once these companies go public, the chance to invest in what he called a “consequential company like this” effectively disappears.

While some investors continue funding companies right up until their IPO in hopes of greater returns, Nvidia’s situation is different. The company is already generating substantial revenue by supplying the chips that power both OpenAI and Anthropic, so there is little need to seek additional returns through further equity investments.

Beyond that straightforward reasoning, several other dynamics may explain Nvidia’s pullback. Industry analysts have frequently pointed out that making large investments in your own biggest customers creates circular and potentially conflicted arrangements, which could lead to negative downstream effects.

When Nvidia first revealed plans last September to invest up to $100 billion in OpenAI, MIT Sloan professor Michael Cusumano described the arrangement to the Financial Times as “kind of a wash,” noting that “Nvidia is investing $100 billion in OpenAI stock and OpenAI is saying they are going to buy $100 billion or more of Nvidia chips.”

This circular logic likely played a role in Nvidia’s decision to scale back its commitment. The investment it finalized just last week, as part of a $110 billion funding round, amounted to $30 billion—far below the $100 billion initially pledged. Huang acknowledged this shift on Wednesday, stating that investing the full amount is “probably not in the cards.”

Some observers have suggested that tensions between the companies could also be a factor, though Huang has dismissed such speculation as “nonsense.” Nevertheless, Nvidia’s relationship with Anthropic has shown signs of strain. Just two months after Nvidia announced a $10 billion investment and a “deep technology partnership” with Anthropic in November, Anthropic CEO Dario Amodei spoke at Davos and, without naming Nvidia directly, compared U.S. chip companies selling high-performance AI processors to approved Chinese customers to “selling nuclear weapons to North Korea.” Those chip companies are Nvidia and AMD.

This week’s developments add further context. Huang’s comments came shortly after the Trump administration blacklisted Anthropic, prohibiting federal agencies and military contractors from using its technology after the company refused to allow its models to be used for autonomous weapons or mass domestic surveillance.

Within hours of that announcement, OpenAI secured its own agreement with the Pentagon—a move that Anthropic has labeled “mendacious” and that public sentiment appears to have viewed similarly. In the following 24 hours, Anthropic’s Claude chatbot surged to the top of Apple’s U.S. App Store, surpassing ChatGPT. At the end of January, Anthropic was not even in the top 100, according to Sensor Tower data.

As a result, Nvidia now holds stakes in two companies that are currently moving in sharply different directions—one newly aligned with the Defense Department, and the other blacklisted by it. Whether Huang anticipated these developments, given Nvidia’s extensive partnership network, is unclear.

However, his stated reason on Wednesday for likely ending future investments—that the IPO window closes the door on such deals—does not fully align with how late-stage private investing typically operates. Given everything that has unfolded recently, it appears more probable that this is an exit from a situation that has become very complicated, very quickly.