Nvidia Pulls Back from OpenAI and Anthropic: The End of an Era?

Published: March 7, 2026

March 7, 20269 min read

The Big Pivot

At the Morgan Stanley Tech, Media and Telecom Conference in San Francisco, Nvidia CEO Jensen Huang signaled that Nvidia's recent investments in OpenAI and Anthropic are likely its final private investments in both companies. The message marks a clear strategy shift for one of the most influential players in AI infrastructure.[1] [2]

From $100 Billion to $30 Billion

In September 2025, Nvidia and OpenAI announced a strategic memorandum involving up to $100 billion and 10 GW of AI infrastructure. Jensen Huang described it as the next breakthrough for artificial intelligence.[3]

By February 2026, Nvidia's finalized participation in OpenAI's $110 billion round came in at $30 billion, with infrastructure plans revised down to 5 GW. The reduction is substantial and has become a focal point in debates about AI capital allocation discipline.[2] [3]

Why Is Nvidia Stepping Back?

Huang's official explanation is simple: both OpenAI and Anthropic are expected to go public in 2026, which would naturally close the private-investment window.[3] [4]

  • Circular investment optics: Critics argue the original structure looked like a loop where Nvidia buys equity and OpenAI buys Nvidia hardware at similar scale, creating concerns about perceived rather than net capital deployment.[2]
  • AI bubble concerns: The extraordinary size of prior announcements intensified market concerns that AI infrastructure funding might be overheating.[3]
  • Tensions with Anthropic leadership: Public remarks by Anthropic CEO Dario Amodei around advanced chip exports were widely interpreted as criticism of Nvidia's position.[2]
  • Political headwinds: Anthropic also faced policy and procurement pressures in Washington, adding complexity to the broader ecosystem narrative.[2]

What This Means for the AI Ecosystem

Nvidia still holds meaningful stakes in OpenAI and Anthropic, and both companies remain deeply dependent on Nvidia chips and compute infrastructure. In practical terms, this looks less like a breakup and more like portfolio risk management before expected IPOs. The partnerships remain strategically significant even as check sizes get smaller.[2] [3]

The Bottom Line

Nvidia's pullback appears to be strategic recalibration, not a collapse in relationships with OpenAI or Anthropic. Huang dismissed suggestions of serious conflict as "nonsense." Capital may be more selective going forward, but the chip-and-infrastructure alliance at the core of the AI stack remains intact.[1] [2]

Sources

More in News

The End of an Era: Alibaba's Qwen Loses Its Architect at a Critical Moment

March 5, 202612 min read

Junyang Lin, the technical lead behind Alibaba's flagship AI platform Qwen, announced his departure on March 3, 2026. Alibaba's shares tumbled as much as 5.3% in Hong Kong as investors reassessed the company's AI ambitions.

AlibabaQwenJunyang LinAI LeadershipOpen Source AIChina AI

Nano Banana 2 Is Here — Google's Most Powerful Image Model Yet

March 3, 20268 min read

Google DeepMind's Nano Banana 2, powered by Gemini 3.1 Flash, brings Pro-level image quality at Flash speed. High-fidelity visuals, subject consistency for up to five characters, precision text rendering, and SynthID watermarking — the future of AI image generation is here.

GoogleNano Banana 2GeminiImage GenerationAIDeepMind