Nvidia Hits the Brakes on OpenAI and Anthropic as Tensions Rise

Nvidia is signaling it may be done writing ever-larger equity checks to the AI startups it supplies, even as it remains the primary company selling the chips that make their ambitions possible. Speaking at the Morgan Stanley Technology, Media and Telecom Conference in San Francisco on March 4, CEO Jensen Huang said Nvidia’s recent investments in OpenAI and Anthropic are likely to be its last in both companies, arguing that once they go public, the window for private investing effectively closes (Loizos; Capoot). The comments landed with extra weight because Nvidia’s relationship with OpenAI has been surrounded by shifting numbers and mounting speculation, moving from a headline-grabbing figure of up to $100 billion discussed last September to a far smaller finalized stake, and now to language suggesting the company is stepping back from further equity exposure (Capoot; “Nvidia Scraps $100B OpenAI Investment Plan”).

The most concrete number today is Nvidia’s $30 billion investment in OpenAI, revealed as part of OpenAI’s massive $110 billion funding round announced the previous Friday (Capoot; “Nvidia Scraps $100B OpenAI Investment Plan”). CNBC reported the round also included a $50 billion commitment from Amazon and a $30 billion commitment from SoftBank, underscoring how concentrated and competitive the AI financing landscape has become (Capoot). Huang said the earlier $100 billion opportunity is “probably not in the cards,” and he tied that directly to OpenAI’s expected IPO toward the end of the year (Capoot). Behind the scenes, Nvidia has also been tempering expectations in its public filings: it disclosed in a quarterly filing in November that the previously announced $100 billion deal might not come to fruition, and it repeated similar caution in February, noting there was “no assurance” it would enter an “investment and partnership agreement” or complete a transaction (Capoot). A Wall Street Journal report in January described the larger agreement as “on ice,” reinforcing the sense that Nvidia was already recalibrating before Huang said it out loud (Capoot).

TechCrunch argues the IPO explanation only partly fits how late-stage investing typically works, and it suggests Nvidia’s pullback may be driven by how complicated the company’s AI ties have become (Loizos). One concern is the circular structure of the biggest infrastructure partnerships. When Nvidia first floated investing up to $100 billion in OpenAI stock, MIT Sloan professor Michael Cusumano described the arrangement as “kind of a wash,” pointing out that OpenAI was also signaling plans to spend comparable sums buying Nvidia’s chips (Loizos). As worries about an AI investment bubble grew, Nvidia’s commitment appears to have shrunk, with the company landing at $30 billion rather than the earlier headline figure (Loizos). Nvidia has also faced persistent rumors about tensions with OpenAI, though Huang dismissed the idea of bad blood as “nonsense,” according to TechCrunch (Loizos).

Strategically, Nvidia can still win without deeper equity stakes because it is paid at the infrastructure layer. CNBC noted OpenAI secured 3 gigawatts of dedicated inference capacity and 2 gigawatts of training capacity on Nvidia’s Vera Rubin systems for AI data centers as part of the broader relationship, tying OpenAI’s near-term computing roadmap to Nvidia hardware (Capoot). At the same time, the industry mix is evolving: AI demand is shifting from training toward inference, which is the fast processing that powers user queries, and that shift has put pressure on Nvidia to deliver new, inference-optimized products. CNBC reported Nvidia is developing a new chip specifically for inference, with OpenAI expected to be one of the largest customers, even as OpenAI also invests in inference-optimized chips from Amazon and uses Google’s TPUs (Capoot). In other words, Nvidia can prioritize selling compute at scale while limiting the financial and reputational risk that comes with being a large shareholder.

Anthropic adds another reason Nvidia might prefer to cap its exposure. Huang said Nvidia’s $10 billion investment in Anthropic will likely be its last as well (Capoot). TechCrunch described Nvidia’s Anthropic relationship as fraught: after Nvidia’s investment, Anthropic CEO Dario Amodei criticized U.S. chip companies selling high-performance AI processors to approved Chinese customers by comparing it to “selling nuclear weapons to North Korea,” without naming Nvidia directly (Loizos). More recently, TechCrunch reported the Trump administration blacklisted Anthropic, barring 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 (Loizos). Within hours of that blacklist, OpenAI announced its own deal with the Pentagon, a move Anthropic labeled “mendacious,” and public attention quickly followed. TechCrunch reported that within 24 hours, Anthropic’s Claude rose to the top of Apple’s U.S. App Store free-app rankings, overtaking ChatGPT, after being outside the top 100 at the end of January, according to Sensor Tower data (Loizos).

Put together, the story looks less like Nvidia losing interest in AI and more like Nvidia choosing a cleaner path: remain the indispensable supplier to everyone building frontier models, but avoid deeper entanglement in the political, reputational, and late-stage financing complexity that now surrounds its biggest customers (Loizos; Capoot; “Nvidia Scraps $100B OpenAI Investment Plan”).

Works Cited

Capoot, Ashley. “Nvidia CEO Huang Says $30 Billion OpenAI Investment ‘Might Be the Last.’” CNBC, 4 Mar. 2026, www.cnbc.com/2026/03/04/nvidia-huang-openai-investment.html. 

Loizos, Connie. “Jensen Huang Says Nvidia Is Pulling Back from OpenAI and Anthropic, but His Explanation Raises More Questions Than It Answers.” TechCrunch, 4 Mar. 2026, techcrunch.com/2026/03/04/jensen-huang-says-nvidia-is-pulling-back-from-openai-and-anthropic-but-his-explanation-raises-more-questions-than-it-answers/. 

“Nvidia Scraps $100B OpenAI Investment Plan.” MSN, 2026, www.msn.com/en-us/news/insight/nvidia-scraps-100b-openai-investment-plan/gm-GM5B6342A2?gemSnapshotKey=GM5B6342A2-snapshot-1&uxmode=ruby. 

Bennett, Bridget. “Jensen Huang, co-founder and chief executive officer of Nvidia Corp., speaks while holding the company’s new GeForce RTX 50 series graphics cards and a Thor Blackwell robotics processor during the 2025 CES event in Las Vegas, Nevada, US, on Monday, Jan. 6, 2025.” Bloomberg, 14 Mar. 2025, www.bloomberg.com/news/articles/2025-03-14/nvidia-ceo-jensen-huang-eyes-next-move-to-outlast-current-ai-boom. 

Ruvic, Dado. “Open AI and Anthropic logos are seen in this illustration created on September 12, 2025.” Reuters, 7 Feb. 2026, www.reuters.com/business/media-telecom/anthropic-buys-super-bowl-ads-slap-openai-selling-ads-chatgpt-2026-02-07/.