Nvidia Invests in OpenAI to Build the World’s Largest AI Cluster

Nvidia has announced an investment of up to 100 billion dollars in OpenAI, as Sam Altman’s company begins the construction of data centers worth hundreds of billions of dollars, based on Nvidia’s artificial intelligence graphics processors. The plan includes systems with a total power of 10 gigawatts, which would be one of the largest AI clusters ever, with an estimated 4–5 million GPUs, which is double what Nvidia delivered last year, according to CNBC.

The first phase of the investment amounts to 10 billion dollars and will be activated upon the completion of the first one-gigawatt data center. Investments will be made gradually, depending on construction progress and market valuations. Nvidia has already felt the impact of the announcement, with its shares jumping nearly four percent, adding about 170 billion dollars to its market value, which now stands at 4.5 trillion.

Nvidia’s CEO Jensen Huang called the project ‘gigantic’ and ‘monumental’, while Altman emphasized that OpenAI must simultaneously develop new products, conduct research, and address unprecedented infrastructure challenges. OpenAI currently has about 700 million weekly users and is valued at 500 billion dollars in the latest secondary round.

Microsoft remains a key partner of OpenAI through Azure and Office integrations, and among the investors are SoftBank and Thrive Capital. Altman points out that both Microsoft and Nvidia are ‘passive’ investors but also the most important strategic partners. The first phase of the project is expected to be operational by 2026, using Nvidia’s new generation of systems named Vera Rubin.

This investment surpasses all previous plans and comes at a time when Nvidia is expanding its portfolio: it has acquired a 5 billion dollar stake in Intel for collaboration on AI processors, invested 700 million in the British startup Nscale, and spent nearly a billion on acquiring the team and licenses of the AI company Enfabrica.

AI Mania Among American Companies

Nvidia’s investment fits into a broader trend where the largest American companies increasingly highlight artificial intelligence, yet many still struggle to explain how it specifically benefits them. An analysis by the Financial Times, based on hundreds of corporate reports and transcripts from S&P 500 companies, shows that AI is mentioned almost everywhere, but often more out of fear of missing out than due to clear business benefits.

Tech giants like Microsoft, Alphabet, Amazon, and Meta have invested hundreds of billions of dollars in infrastructure for large language models, while companies outside Silicon Valley, from Coca-Cola to Lululemon, emphasize caution in their reports. They most frequently mention risks such as cybersecurity, legal challenges, and potential implementation failures.

Certain industries, however, offer clearer examples. Paycom, a payroll service provider, states that AI is an important differentiator in attracting clients. Huntington Ingalls uses it for military decisions, Zoetis for accelerated medical testing for horses, and Dover Corporation for tracking and processing vehicles damaged by hail. However, despite innovations, the stock growth of these companies does not deviate from the broader market.

According to data, 374 of the 500 largest companies mentioned AI on conference calls in the past year, and as much as 87 percent of those mentions were exclusively positive. However, official regulatory reports (SEC 10-K) are much more cautious and often highlight potential negative consequences. The most frequently expressed risk is cybersecurity – more than half of the companies cited it in 2024. Match Group, the owner of online dating apps, warned that AI could lead to incidents involving users’ personal data, while Microsoft noted that inadequate application of artificial intelligence could cause harm to users and society.

Another major issue is the failure of AI initiatives. A study by Microsoft and MIT Media Lab found that 95 projects of generative artificial intelligence in work environments fail, most often because the tools lack long-term memory or sufficient adaptability. Additionally, the number of legal and regulatory threats is increasing. Meta warned that it could incur significant costs due to lawsuits related to the use of copyrighted materials to train its models. PepsiCo cited similar risks due to potential unauthorized use of third-party technologies or content.

The FT analysis shows that companies are much clearer when discussing risks than when addressing the specific benefits of artificial intelligence. ‘There is no guarantee that the use of artificial intelligence will improve our products or business, nor that it will be beneficial for our efficiency or profitability,’ Meta wrote in its report.