[News] NVIDIA’s Top 2 Customers Jump to 39% of Sales from 25% Last Year, Raising Client Dependence Concerns

While the public has growing concerns after NVIDIA’s earnings call on H20 related issues and the uncertainties in the Chinese market, CNBC, citing NVIDIA’s 10-Q filing, points out overly client concentration may be another problem. According to the report, two customers accounted for 39% of NVIDIA’s revenue in the quarter ending July.

According to NVIDIA’s SEC filing, China’s share of the company’s sales has fallen to 5.9%, down from 12.5% in the April quarter, trailing the U.S. (50.2%), Singapore (21.7%), and Taiwan (18.2%).

Meanwhile, CNBC highlights that “Customer A” accounted for 23% of total revenue in the July quarter, with “Customer B” adding 16%—up sharply from 14% and 11% for the top two customers a year earlier.

Notably, NVIDIA’s latest disclosure has reignited concerns that its fast growth may depend heavily on a handful of major cloud providers, including Microsoft, Amazon, Google, and Oracle, as per CNBC. This aligns with CFO Colette Kress’ remarks, as Reuters reported she said roughly half of NVIDIA’s $41 billion data center revenue last quarter came from these cloud giants—a critical segment that made up 88% of total Q2 revenue.

Who Are NVIDIA’s Mysterious Top Customers?

The disclosure has stirred market speculation about the identities of NVIDIA’s top customers—but they may not be the cloud providers most expect. According to the company, those who buy products directly from NVIDIA are classified as direct customers, including add-in board makers, distributors, ODMs, OEMs, and system integrators.

CNBC notes that direct customers can range from original design manufacturers (ODMs) or original equipment manufacturers (OEMs) like Foxconn and Quanta to distributors and system integrators such as Dell. Both “Customer A” and “Customer B” fall into this direct customer category.

As NVIDIA explains, indirect customers are those that buy products through its direct customers. These include cloud service providers (CSPs), consumer internet firms, enterprises, and public sector organizations, the company notes.

However, as CNBC suggests, pinpointing whether any of these cloud providers are the undisclosed Customers A and B is tricky, partly because NVIDIA has flexibility in defining direct versus indirect customers. According to NVIDIA’s filing, one indirect customer—primarily purchasing through Direct Customer A—is estimated to represent 10% or more of total revenue and attributable to the Compute & Networking segment.

Regardless of who its top clients turn out to be, NVIDIA is poised to take the bulk of spending in the cloud-driven AI data center boom. CEO Jensen Huang told Reuters he expects $3–4 trillion in AI infrastructure investment by the end of the decade, noting that for a data center costing up to $60 billion, NVIDIA could capture around $35 billion.

Read more

• [News] NVIDIA Growth Risk Looms as Revenue Expansion Hits Two-Year Low, China Hurdles Cloud Outlook
• [News] NVIDIA Reportedly Eyes Small-Scale HBM Base Die Production in 2027: Rattling Memory, Chip Markets

(Photo credit: NVIDIA’s X)