Semiconductor giant NVIDIA (NASDAQ:NVDA) plans to invest up to US$100 billion in OpenAI to build what executives are calling the largest artificial intelligence (AI) infrastructure project in history.
The companies said on Monday (September 22) that OpenAI will deploy NVIDIA’s systems on a scale requiring 10 gigawatts of power, equal to 4 million to 5 million of the chipmaker’s graphics processing units (GPUs).
NVIDIA CEO Jensen Huang described the effort as “monumental in size,” noting in an interview with CNBC that it represents roughly double the volume of GPUs shipped last year.
“NVIDIA and OpenAI have pushed each other for a decade, from the first DGX supercomputer to the breakthrough of ChatGPT,” Huang said in the announcement. Company shares rose nearly 4 percent on the day of the announcement, adding close to US$170 billion in market value and lifting NVIDIA’s market capitalization near US$4.5 trillion.
“Everything starts with compute,” Sam Altman, co-founder and CEO of OpenAI, added.
“Compute infrastructure will be the basis for the economy of the future, and we will utilize what we’re building with NVIDIA to both create new AI breakthroughs and empower people and businesses with them at scale.”
NVIDIA will deploy the funding to OpenAI progressively as data center capacity comes online. An initial US$10 billion tranche is tied to the completion of the first gigawatt, with that milestone scheduled for 2026.
The move builds on investor enthusiasm after NVIDIA reported record revenue of US$46.7 billion in its latest fiscal quarter, up 56 percent from a year earlier. Its net income jumped 59 percent to US$26.42 billion, with adjusted earnings per share of US$1.05 topping Wall Street forecasts.
The results marked the company’s ninth straight quarter of year-on-year revenue growth above 50 percent.
The chipmaker has been on a deal-making spree. In recent weeks, it disclosed a US$5 billion stake in Intel (NASDAQ:INTC) tied to joint AI processor development, and a nearly US$700 million investment in UK data center startup Nscale. It also announced the US$900 million acquisition of staff and technology from startup Enfabrica.
For OpenAI, the NVIDIA commitment addresses the enormous computing power needed to support its research and rapidly growing user base. According to the company, it now has 700 million weekly active users, an amount that is straining its existing infrastructure even as demand for more advanced models grows.
Industry analysts estimate that building a single gigawatt of AI data center capacity costs US$50 billion to US$60 billion, with roughly US$35 billion of that tied to NVIDIA chips and systems.
Despite this week’s deal, OpenAI has not limited itself to NVIDIA hardware.
The company has reportedly partnered with Broadcom (NASDAQ:AVGO) to develop custom AI chips, with sources identifying OpenAI as the unnamed customer behind a US$10 billion order disclosed by Broadcom CEO Hock Tan. Both firms declined to comment, but analysts see the deal as a bid to reduce reliance on NVIDIA and ease GPU shortages.
NVIDIA’s investment in OpenAI will complement the lab’s existing ties with Microsoft (NASDAQ:MSFT), Oracle (NYSE:ORCL) and the Stargate project. Microsoft remains one of OpenAI’s earliest and most important backers, and has integrated the company’s models into its Azure cloud and Office products.
Securities Disclosure: I, Giann Liguid, hold no direct investment interest in any company mentioned in this article.