Nvidia CEO denies rumors of issues in OpenAI deal

Nvidia CEO Jensen Huang said that the company’s planned investment in OpenAI remains on track, despite recent reports suggesting potential issues.

Huang told CNBC that there is no controversy and emphasized the company’s continued support for OpenAI.

In September, Nvidia and OpenAI announced a letter of intent for Nvidia to invest up to US$100 billion in multiple rounds, with OpenAI planning to build AI infrastructure using Nvidia’s chips.

An SEC filing in November indicated the deal had not been finalized, leading to speculation that the agreement was delayed.

Huang confirmed Nvidia would join OpenAI’s next funding round, which CNBC reported could raise as much as US$100 billion.

OpenAI has used Nvidia GPUs for its AI models since inception, though recent statements from OpenAI’s Sam Altman highlighted chip shortages and the company’s efforts to secure additional hardware from competitors like AMD and Broadcom.

🔗 Source: CNBC

🧠 Food for thought

Implications, context, and why it matters.

Nvidia’s up-to-US$100 billion investment plan in OpenAI remains at the non-binding letter-of-intent stage

The September agreement is a letter of intent (LOI). Nvidia said it intends to invest up to US$100 billion in OpenAI across multiple rounds as new NVIDIA systems are deployed, with terms to be set later 1.
Two months later, Nvidia executive vice president (EVP) and chief financial officer (CFO) Colette Kress said Nvidia still had not signed a definitive agreement with OpenAI 2.
Commentators have called the setup “circular financing,” meaning Nvidia puts in money and OpenAI uses it to buy Nvidia GPUs, though Nvidia’s LOI statement did not use that framing 3. Others have warned vendor financing could make organic demand look higher, though Nvidia has not described the deal that way 4.

The plan’s execution hinges on power availability and supply-chain capacity

Nvidia and OpenAI said the partnership aims to deploy at least 10 gigawatts of NVIDIA systems. That scale would need large amounts of electricity, which could pressure power grids 1.
The rollout could run into supply constraints, including limits in advanced chip packaging plus High Bandwidth Memory (HBM), which industry sources link to long lead times 5.
Nvidia has flagged customer-concentration risk. In one period it cited, two customers made up 23% and 16% of quarterly revenue, which ties results to a small set of large buyers 6.
Analysts have compared the surge in AI infrastructure spending with earlier build cycles, including the dot-com era telecom overbuild 4.