On September 22, NVIDIA and OpenAI jointly declared their intention to forge a groundbreaking strategic partnership, as evidenced by a signed letter of intent.

As per the agreement, OpenAI is set to integrate at least 10 gigawatts of NVIDIA's cutting-edge systems into its next-gen AI infrastructure. This move is aimed at propelling the training and operational capabilities of its forthcoming models, thereby hastening the advent of superintelligence.

NVIDIA has earmarked an investment of up to $100 billion to facilitate extensive deployments, encompassing data center construction and power supply enhancements, throughout the system implementation phase with OpenAI. The inaugural phase of this ambitious project is slated for launch in the latter half of 2026, leveraging NVIDIA's state-of-the-art Vera Rubin platform.

Jensen Huang, NVIDIA's founder and CEO, remarked, "Since the inception of the first DGX supercomputer to the groundbreaking emergence of ChatGPT, NVIDIA and OpenAI have been mutually propelling each other forward for a decade. This investment and infrastructure synergy signify a new milestone for us—harnessing 10 gigawatts of power to ignite the next era of intelligence."

In a CNBC interview, Huang further elaborated that NVIDIA would underpin the construction of the largest data center ever conceived, resulting in the sale of 4 to 5 million NVIDIA chips to OpenAI.

Sam Altman, co-founder and CEO of OpenAI, also commented, "Computing is the cornerstone of everything. The computing infrastructure will lay the groundwork for the future economy. We intend to capitalize on the infrastructure developed in tandem with NVIDIA to pioneer new AI breakthroughs and empower individuals and enterprises on a grand scale."

OpenAI will work hand-in-hand with NVIDIA to co-optimize the development trajectories of OpenAI's models and infrastructure software, alongside NVIDIA's hardware and software offerings. This collaborative effort also complements the in-depth work undertaken by OpenAI and NVIDIA in conjunction with their vast partner ecosystem, including Microsoft, Oracle, SoftBank, and Stargate collaborators.

NVIDIA and OpenAI are eagerly anticipating the finalization of the new strategic cooperation phase's specifics in the upcoming weeks.

NVIDIA's investment in OpenAI aligns with a broader industry trend, where tech firms are collectively investing hundreds of billions of dollars globally in AI data center projects. Data indicates that OpenAI, Amazon, Google, Meta, and Microsoft are poised to invest a cumulative total exceeding $325 billion in these facilities by the year's end.

With annual profits soaring into the billions, these tech behemoths can easily finance data center construction projects using their cash reserves. However, emerging smaller entities like OpenAI, as they embark on similar ventures, are compelled to secure billions in funding or loans.

Numerous experts harbor concerns that if AI technology adoption does not materialize as swiftly as these companies anticipate, such profligate spending could jeopardize their financial stability. Many firms may find themselves saddled with substantial debts, lacking adequate sales to offset expenditures.

According to The New York Times, in January of this year, OpenAI, in collaboration with partners SoftBank and cloud computing titan Oracle, along with President Trump, unveiled plans to invest up to $500 billion in constructing new data center infrastructure across the United States. Dubbed 'Stargate,' this initiative has already seen the establishment of a data center in Texas, with further facilities planned for other U.S. regions.

Moreover, building on an agreement reached between the Trump administration and the United Arab Emirates, OpenAI intends to set up a computing hub in the UAE. This data center forms part of a joint venture involving Oracle, SoftBank, and the UAE's AI firm G42.

Data reveals that NVIDIA's stock price surged by 4.4%, reaching a new intraday peak, while data center constructor Oracle's stock price climbed by approximately 6%.

As per Reuters, OpenAI had previously partnered with design firm Broadcom to craft a bespoke chip. Following this announcement, Broadcom's stock price dipped by 0.8%.

Reuters highlighted that the magnitude of NVIDIA's latest commitment might invite antitrust scrutiny. The U.S. Department of Justice and the Federal Trade Commission struck a deal in mid-2024 to probe Microsoft, OpenAI, and NVIDIA. Nevertheless, Trump did not regard industry competition as a pressing concern.

Earlier this month, OpenAI and its principal backer, Microsoft, also disclosed a non-binding agreement to restructure OpenAI as a for-profit entity, signaling further governance transformations.

Andre Barlow, an antitrust attorney at Doyle, Barlow & Mazard, opined, "This transaction could alter the economic incentives for NVIDIA and OpenAI, potentially solidifying NVIDIA's dominance in the chip sector through OpenAI's software leadership. This could pose challenges for NVIDIA's chip sector rivals (like AMD) or OpenAI's model sector competitors in scaling their operations."

References