Two months prior, Nvidia CEO Jensen Huang and OpenAI CEO Sam Altman gathered in San Jose, California, to unveil a significant agreement between the two frontrunners in artificial intelligence.
Nvidia is set to invest $100 billion over several years, starting in 2026, as OpenAI’s AI supercomputing facilities become operational, as announced by the duo. The specifics regarding the timing of the infrastructure development and the expenses for each data center were not revealed.
However, in Nvidia’s quarterly financial report released on Wednesday, the chip manufacturer reminded stakeholders of the distinction between an announcement and a binding contract.
“There is no guarantee that definitive agreements will be made regarding the OpenAI opportunity or other prospective investments, nor that any investment will proceed on anticipated terms,” Nvidia stated in the risk factors section of its quarterly documentation.
Nvidia has engaged in significant investments lately, deploying its growing cash reserves to support enterprises purchasing its graphics processing units, or GPUs. In addition to the OpenAI agreement, Nvidia on Wednesday emphasized its $5 billion commitment to invest in Intel during the fiscal quarter and its deal announced this week to invest up to $10 billion in Anthropic.
An OpenAI representative didn’t provide a statement but referred to Huang’s remarks during the call, including his characterization of OpenAI as a “once-in-a-generation company” and his expectation that the investment will “yield extraordinary returns.”
“There is no assurance that any investment will be realized on anticipated terms, if at all,” Nvidia reiterated.
The primary distinction with OpenAI lies in the magnitude of the proposed investment and the benchmarks that must be satisfied for the total funding to be realized. A source told CNBC at the time of the announcement that an initial $10 billion would soon be accessible to OpenAI to assist the organization in achieving its first gigawatt of operational capacity.
Altman recently shared that OpenAI anticipates concluding the year with a $20 billion annualized revenue run rate, a remarkable figure considering its flagship ChatGPT product is merely three years old. Altman mentioned that the organization aims to achieve hundreds of billions of dollars in revenue by 2030. Nonetheless, that projected amount does not sufficiently cover the company’s costs.
Overall, OpenAI has declared approximately $1.4 trillion in infrastructure expenditures with multiple partners as it strives to expand its AI models and services. To attain this, the organization relies on external funding.
Despite the ambiguity surrounding the September agreement, Nvidia leaders remain optimistic about their collaboration with OpenAI. During Nvidia’s earnings call, after the chipmaker reported a strong revenue and earnings performance along with better-than-anticipated guidance, CFO Colette Kress praised OpenAI’s growth.
“OpenAI recently indicated that their weekly user count has risen to 800 million, the number of enterprise clients has reached 1 million, and their gross margins are robust,” Kress noted. She mentioned that the two firms are “developing a strategic partnership” and that Nvidia is “dedicated to aiding them in constructing and deploying at least 10 gigawatts of AI data centers.”
Huang remarked, “Everything that OpenAI engages in operates on Nvidia today.”
An OpenAI representative didn’t offer a comment but cited Huang’s observations on the call, including his description of OpenAI as a “once-in-a-generation company” and his belief that the investment will “translate to extraordinary returns.”
It is undeniable that as OpenAI develops data centers, it will continue to invest in Nvidia’s chips. However, OpenAI has also formed a partnership with Nvidia competitor Advanced Micro Devices, agreeing last month to deploy 6 gigawatts of AMD’s Instinct GPUs over several years and multiple generations of hardware, commencing in the latter half of next year.
The AMD agreement features one vital aspect that Nvidia’s agreement lacks: signatures.
As part of the arrangement, the company has granted OpenAI a warrant for up to 160 million shares of the chipmaker’s common stock, with milestones for vesting linked to deployment volume and AMD’s stock price. That deal was signed on October 5, by AMD CFO Jean Hu and OpenAI CFO Sarah Friar.
— CNBC’s MacKenzie Sigalos and Ashley Capoot contributed to this report.
WATCH: Nvidia revenue is a more significant narrative than gross margins looking ahead
