Nvidia Reveals: Two Mystery Customers Account for About 40% of Revenue
Regulatory filing shows nearly 40% of Nvidia’s second-quarter revenue came from just two unnamed clients, highlighting concentrated sales tied to AI and data center demand
Nearly 40% of Nvidia’s revenue in the second quarter of its fiscal year came from only two customers, according to a filing submitted to the U.S. Securities and Exchange Commission.
The company referred to them only as “Customer A” and “Customer B,” without disclosing their names.
The filing stated that Customer A was responsible for 23% of Nvidia’s revenue and Customer B for 16%.
This marks an increase compared with last year, when the two largest clients accounted for 14% and 11% respectively.
The document clarified that these were “direct customers,” such as original equipment manufacturers (OEMs), integrators, or distributors—companies that purchase Nvidia chips to build full systems, which are then sold to data centers, cloud service providers, and end users.
Indirect customers, including major cloud operators and consumer internet firms, typically obtain Nvidia’s chips through these direct buyers.
Nvidia explained it can estimate sales to indirect customers from internal data and purchase orders.
The filing noted that two indirect customers each accounted for more than 10% of revenue, largely via purchases from Customers A and B.
In the first half of the fiscal year, these two clients represented 20% and 15% of Nvidia’s total revenue.
Four other customers followed with shares of 14%, 11%, 11%, and 10% respectively in the second quarter.
The company also disclosed that an artificial intelligence research and development firm made a “significant” contribution to revenue through both direct and indirect channels.
Last week Nvidia reported record revenue of 46.7 billion dollars, a 56% increase from a year earlier, driven largely by its data center business linked to artificial intelligence demand.
While the identities of Customers A and B remain undisclosed, analysts suggest that major cloud service providers—including Microsoft, Meta, Amazon, Google, and Oracle—are likely among the indirect clients sourcing chips through OEMs or distributors.
According to industry reports, large cloud operators such as Microsoft are estimated to account for around 50% of Nvidia’s data center revenue, and cloud providers overall contributed approximately 88% of total company revenue.
Analysts describe Nvidia’s sales mix as highly concentrated, with Microsoft, Google, and Meta generally considered the largest buyers, though figures fluctuate from quarter to quarter.
Regardless of whether sales are recorded as direct or through distributors, demand stems primarily from the surge in artificial intelligence and data center investment.
Cloud leaders including Amazon, Google, Meta, and Microsoft collectively spent an estimated 350 billion dollars in capital expenditures, with current quarterly figures suggesting they represent about 70% of Nvidia’s revenue.
In the most recent period, Amazon invested 32 billion dollars, Google 22 billion, and Meta and Microsoft 17 billion each.
Some analysts caution that revenue concentration among a small group of customers poses risks, even as these firms possess abundant cash flow and are committed to massive spending on data infrastructure in the next two years.
Others argue it is not a weakness but rather inherent to the business, since only a handful of companies globally can allocate 100 billion dollars annually to such investments.
Nvidia’s CEO Jensen Huang said the company projects between 3 trillion and 5 trillion dollars in global artificial intelligence infrastructure spending by the end of the decade.
He estimated Nvidia could capture about 70% of the total cost of an AI-focused data center through sales of its graphics processing units and other chips.
Huang also noted that all major cloud companies, including Meta and Microsoft, have committed to spending 600 billion dollars this year, alongside additional firms joining the race to build AI-ready data centers.
The company referred to them only as “Customer A” and “Customer B,” without disclosing their names.
The filing stated that Customer A was responsible for 23% of Nvidia’s revenue and Customer B for 16%.
This marks an increase compared with last year, when the two largest clients accounted for 14% and 11% respectively.
The document clarified that these were “direct customers,” such as original equipment manufacturers (OEMs), integrators, or distributors—companies that purchase Nvidia chips to build full systems, which are then sold to data centers, cloud service providers, and end users.
Indirect customers, including major cloud operators and consumer internet firms, typically obtain Nvidia’s chips through these direct buyers.
Nvidia explained it can estimate sales to indirect customers from internal data and purchase orders.
The filing noted that two indirect customers each accounted for more than 10% of revenue, largely via purchases from Customers A and B.
In the first half of the fiscal year, these two clients represented 20% and 15% of Nvidia’s total revenue.
Four other customers followed with shares of 14%, 11%, 11%, and 10% respectively in the second quarter.
The company also disclosed that an artificial intelligence research and development firm made a “significant” contribution to revenue through both direct and indirect channels.
Last week Nvidia reported record revenue of 46.7 billion dollars, a 56% increase from a year earlier, driven largely by its data center business linked to artificial intelligence demand.
While the identities of Customers A and B remain undisclosed, analysts suggest that major cloud service providers—including Microsoft, Meta, Amazon, Google, and Oracle—are likely among the indirect clients sourcing chips through OEMs or distributors.
According to industry reports, large cloud operators such as Microsoft are estimated to account for around 50% of Nvidia’s data center revenue, and cloud providers overall contributed approximately 88% of total company revenue.
Analysts describe Nvidia’s sales mix as highly concentrated, with Microsoft, Google, and Meta generally considered the largest buyers, though figures fluctuate from quarter to quarter.
Regardless of whether sales are recorded as direct or through distributors, demand stems primarily from the surge in artificial intelligence and data center investment.
Cloud leaders including Amazon, Google, Meta, and Microsoft collectively spent an estimated 350 billion dollars in capital expenditures, with current quarterly figures suggesting they represent about 70% of Nvidia’s revenue.
In the most recent period, Amazon invested 32 billion dollars, Google 22 billion, and Meta and Microsoft 17 billion each.
Some analysts caution that revenue concentration among a small group of customers poses risks, even as these firms possess abundant cash flow and are committed to massive spending on data infrastructure in the next two years.
Others argue it is not a weakness but rather inherent to the business, since only a handful of companies globally can allocate 100 billion dollars annually to such investments.
Nvidia’s CEO Jensen Huang said the company projects between 3 trillion and 5 trillion dollars in global artificial intelligence infrastructure spending by the end of the decade.
He estimated Nvidia could capture about 70% of the total cost of an AI-focused data center through sales of its graphics processing units and other chips.
Huang also noted that all major cloud companies, including Meta and Microsoft, have committed to spending 600 billion dollars this year, alongside additional firms joining the race to build AI-ready data centers.