Nvidia’s top two mystery customers made up 39% of its Q2 revenue
Two Nvidia Customers made up 39% of the Nvidia income in July quarter, the company announced a financially filing On Wednesday, to increase concerns about the client’s concentration of the chip manufacturer.
“Customer A” has made 23% of the total income, and “Customer B” constituted 16% of the total income according to the Company’s Securities and Stock Exchange Commission.
According to the file, in the previous quarter, NVIDIA is higher than the same quarter, where the two best customers make up 14% and 11% of sales.
The company publishes information about its best customers every three months, but this week’s statement provides a renewed debate on whether NVIDIA’s explosive growth will be directed by a handful of large cloud provider. Microsoft– Amazon– Google And Prophecy.
Nvidia Finance Chief Colette Kress said on Wednesday that “large cloud service providers” constitute approximately 50% of the company’s data center income. This is important because the data center job constitutes 88% of NVIDIA’s total income in the second quarter.
“We have experienced a significant amount of income from a limited number of customers and this tendency may continue.”
Analysts are increasingly looking for their commitments to spend these cloud capital expenditures to model NVIDIA’s future growth.
“Unless it has increased clarity in 2026, more earnings are limited to upward revision or stock price catalyst. [cloud service provider] Capex expectations, “he wrote HSBC analyst Frank Lee with a note on Thursday.
However, NVIDIA’s customer A and Customer B are not necessarily cloud providers. A little mystery and a NVIDIA representative refused to share the identities of A and Customer B.
Nvidia says that there are both “direct customers” and “indirect customers” in the file. Customer A and Customer B are listed as “direct customers”.
Direct customers are not the end users of Nvidia’s chips. They are then the companies that buy chips to convert them to data centers, cloud providers and end users to complete systems or circuit cards they sell to end users. Some of these direct customers are original design manufacturers or original equipment manufacturers such as Foxconn or Quanta. Others are distributors or system integrators like Dell.
In the meantime, indirect customers usually include cloud service providers, internet companies and businesses that buy system directly from NVIDIA’s customers. NVIDIA, income, only to indirect customers, purchasing orders and internal sales data can be predicted, he said.
NVIDIA’s mysterious customers of any of these cloud providers are difficult to decipher, because the chip manufacturer moves directly in the definitions of their direct and indirect customers.
For example, NVIDIA wrote that some direct customers purchase chips to create systems for their use.
In addition, NVIDIA said that two of the indirect customers constitute more than 10% of the total income of each of them and that they purchased the system through A and B customers.
Nvidia, who contributed more to the mystery of everything, said that a “AI research and development company” contributes to a “meaningful” income through directly and indirect customers.
On Wednesday, the NVIDIA said that the demand for the company’s AI systems, not only among cloud providers, but also remained higher with morely adjusted services for AI, including System Purchasing Systems for AI and other customer types, including “Neoelouds”. Nvidia also listed foreign governments and said that this year would record $ 20 billion for “Egemen AI”. Kres, in his statement to analysts about a call for earnings, all of these product categories contribute to Nvidia’s increase in revenue.
Nvidia CEO Jensen Huang said that the company’s AI infrastructure is a new estimation of $ 3 in the AI infrastructure by the end of a decade. He said that a 50 billion -dollar AI -oriented data center, which he sells not only for graphic processing units, but also for other chips, he said he could get approximately 70% of his total cost.
According to Huang, Huang said that Huang is a logical target for the next five years. He also said that new customer types such as businesses or overseas cloud providers have participated in the accumulation.
“As you know, while the AI revolution is full evaporation, only the first four Hyperscaler Capex has doubled in two years.” He said.
WRISTWATCH: Nvidia’s risk of concentration addiction
