OpenAI execs, including Sam Altman, have posted public defenses of Nvidia following a report from Reuters that claimed the generative AI company is unsatisfied with some Nvidia chips and has been looking for alternatives since last year.

Per the report, OpenAI is increasingly looking to deploy chips that can support specific AI inferencing tasks. Seven sources cited by the news outlet said that the company is not satisfied with the speed at which Nvidia hardware is able to provide ChatGPT users with answers to certain queries, including those related to software development and agentic AI.

As a result, one of the sources said OpenAI has been looking for alternative hardware that would ultimately be able to support around 10 percent of its future inferencing needs, placing a particular focus on companies offering hardware with large amounts of embedded SRAM (static random-access memory).

Although more expensive than other types of memory, such as DRAM (dynamic random-access memory), SRAM is faster and therefore allows AI applications such as chatbots to surface answers more quickly.

The sources claim that these concerns from OpenAI led to Nvidia approaching Cerebras and Groq regarding potential acquisitions. In December 2025, Nvidia announced plans to license technology from Groq and hire its leadership team. Though the deal is not an outright acquisition, it is reportedly worth up to $20 billion.

The following month, Cerebras struck a multi-year deal with OpenAI, reportedly worth more than $10 billion for the provision of around 750MW of computing power, set to be delivered through 2028.

Following the publication of the report from Reuters, OpenAI CEO Sam Altman took to social media platform X to say that the startup loves working with Nvidia, “and they make the best chips in the world,” adding: “We hope to be a gigantic customer for a very long time.”

He concluded the post by saying: “I don't get where all this insanity is coming from.”

Late in the day (January 2), Sachin Katti, head of compute infrastructure at OpenAI, posted a longer statement on X, defending the company’s partnership with Nvidia.

“Our partnership with Nvidia is foundational,” he wrote. “Nvidia is our most important partner for both training and inference, and our entire compute fleet runs on Nvidia GPUs. This is not a vendor relationship. It is deep, ongoing co-design. We build systems together, and our frontier models are the product of multi-year hardware and model engineering done side by side.”

He went on to say that as inferencing demand continues to grow “exponentially,” Nvidia “continues to set the bar for performance, efficiency, and reliability for both training and inference.” However, he said that in order to deploy “more broadly, and support the explosion of real-world use cases without sacrificing performance or reliability,” OpenAI was also deliberately expanding its ecosystem through partnerships with Cerebras, AMD, and Broadcom.

“The outcome is simple and durable: infrastructure that can carry frontier capability all the way into production, at [a] global scale,” Katti wrote.

Katti joined OpenAI in November 2025, six months after accepting the role as Intel’s chief technology and AI officer.

Nvidia CEO walks back $100bn OpenAI investment pledge

Separately, Nvidia CEO Jensen Huang has faced questions from the media this week after he downplayed a previous pledge to invest up to $100 billion in OpenAI. A letter of intent was signed by the two companies in September 2025, under the terms of which Nvidia said it would deploy at least 10GW of its systems in OpenAI’s data centers, whilst also investing up to $100 billion in OpenAI as the new systems come online.

At the time, Huang called it “the largest computing project in history," but the company's CFO said that the deal had not passed the letter of intent stage as of December.

Earlier this week, the Wall Street Journal reported that OpenAI had expected the pledge to be finalized within weeks, but progress has yet to be made, with Huang going on to publicly state: “We never said we were going to invest $100 billion in one round. They invited us to invest up to $100 billion. And, of course, we were very happy and honored."

The WSJ further claimed that Huang has also privately criticised a lack of discipline in OpenAI's business approach and shared concern about the competition it faces from Google, Anthropic, and others.

Asked about the article, Huang said it was “nonsense” to say he was unhappy with OpenAI, going on to add that Nvidia was still planning to invest ‍a “great deal of money” in OpenAI, calling it “probably the largest investment we’ve ever made.”