Three Mysterious Buyers Spent Billions on Nvidia’s AI Chips

AI microchip manufacturer Nvidia, the world’s most valuable company by market capitalization, continues to rely heavily on a handful of anonymous buyers who collectively contribute tens of billions of dollars in revenue. The elite trio of buyers, particularly deep-pocketed, individually purchased between $10 billion and $11 billion worth of goods and services in the first nine months ending in late October. Fortunately for Nvidia investors, this is not expected to change anytime soon. Mandeep Singh, global head of technology research for Bloomberg Intelligence, says he believes the forecast of founder and CEO Jensen Huang that spending will not cease, as reported by Fortune.

Nvidia Still Has Limited Supply

It is very unusual for a company to have such a concentration of risk among a handful of buyers – let alone one that is poised to become worth an astronomical sum of four trillion dollars. Singh told Fortune that the anonymous buyers likely include Microsoft, Meta, and possibly Super Micro. However, Nvidia has declined to comment on speculation. Nvidia refers to them only as buyers A, B, and C, and all have purchased goods and services totaling $12.6 billion. This accounted for more than a third of Nvidia’s total of $35.1 billion recorded for the fiscal third quarter ending in October.

Their share is also divided into equal parts, with each amounting to 12 percent, suggesting that they were likely receiving the maximum amount of chips allocated to them, rather than as many as they would ideally want. Founder and CEO Jensen Huang emphasized that his company has a limited supply. Nvidia cannot produce more chips, as it has outsourced the wholesale production of its industry-leading AI microchips to Taiwan’s TSMC and does not have its own manufacturing facilities.

Nvidia keeps their identity a business secret for competitive reasons—undoubtedly these buyers would not want their investors, employees, critics, activists, and rivals to see exactly how much money they are spending on Nvidia’s chips. For example, one party labeled as “Buyer A” purchased about $4.2 billion in goods and services during the last fiscal quarterly period. However, it seems that in the past this amount was less, as it does not exceed the 10 percent threshold in the first nine months overall.

Long-Term Risks for Nvidia

Oracle recently announced plans to build a zettascale data center with more than 131,000 state-of-the-art Nvidia Blackwell AI chips. It is estimated that the electricity required to run such a massive computing cluster would be equivalent to the output capacity of nearly two dozen nuclear power plants.

Bloomberg Intelligence analyst Singh actually sees only a few long-term risks for Nvidia. First, some large buyers are likely to eventually reduce orders, diluting its market share. One such likely candidate is Alphabet, which has its own training chips called TPUs.

Second, its dominance in training is not aligned with inference, which powers generative AI models after they have already been trained. Here, the technical requirements are not nearly as high-end, meaning there is much greater competition not only from rivals like AMD but also from companies with their own custom models like Tesla.