Nvidia’s latest SEC filing reveals a fascinating dynamic: two major clients accounted for a whopping 39% of the chipmaker’s revenue in the July quarter. This concentration, disclosed on Wednesday, is fueling discussions about the sustainability and breadth of Nvidia’s monumental growth.
According to the filing, “Customer A” comprised 23% of total revenue, while “Customer B” accounted for 16%. This figure is a significant jump from the same quarter last year, where the top two customers represented 14% and 11% of sales, respectively.
While Nvidia routinely discloses information about its major customers, the increased concentration is prompting analysts to examine whether the company’s explosive growth is overly reliant on a relatively small group of hyperscale cloud providers like Microsoft, Amazon, Google, and Oracle.
Nvidia’s CFO, Colette Kress, noted that “large cloud service providers” contribute approximately 50% of the company’s data center revenue. Given data centers represent a staggering 88% of Nvidia’s overall Q2 revenue, this is a critical piece of the puzzle.
The company itself acknowledges the potential risk: “We have experienced periods where we receive a significant amount of our revenue from a limited number of customers, and this trend may continue,” Nvidia stated in the filing, a prudent caveat for investors.
This dependence is leading analysts to meticulously track cloud capital expenditure commitments to gauge Nvidia’s future trajectory. HSBC analyst Frank Lee, who holds a “hold” rating on the stock, observed, “We see limited room for further earnings upside revision or share price catalyst in the near-term unless we have increasing clarity over upside in 2026 [cloud service provider] capex expectations.” In essence, cloud spending is increasingly seen as a bellwether for Nvidia’s continued ascent.
Here’s where the mystery deepens: it’s not definitively clear if Customer A and Customer B are cloud providers. An Nvidia representative remained tight-lipped, declining to identify the specific companies. This ambiguity adds another layer to the analysis.
Nvidia distinguishes between “direct customers” and “indirect customers,” with Customers A and B classified as “direct.” These are not the end users of Nvidia’s chips. Instead, they are companies that purchase the chips to integrate them into complete systems or circuit boards. These systems are then sold to data centers, cloud providers, and end-users. This category includes original design manufacturers (ODMs) and original equipment manufacturers (OEMs) like Foxconn and Quanta, as well as distributors and system integrators such as Dell.
Indirect customers, by contrast, encompass cloud service providers, internet giants, and large enterprises that typically acquire systems from Nvidia’s direct clientele. Nvidia estimates revenue from indirect customers based on purchase orders and internal sales data, adding a level of inherent estimation to their figures.
The blurring of lines between direct and indirect customer roles further complicates matters. Nvidia acknowledges in its filing that certain direct customers acquire chips to build systems for their internal use, muddying the waters in attributing revenue to specific end users.
Adding another intriguing detail, Nvidia disclosed that two of its indirect customers each accounted for over 10% of its total revenue, primarily through purchases made via Customers A and B. This hints at significant downstream influence from specific players.
Further adding to the intrigue, Nvidia highlighted that an “AI research and development company” contributed a “meaningful” amount of revenue through both direct and indirect channels. This suggests a diversifying customer base extending beyond the traditional hyperscalers.
During Wednesday’s earnings call, Nvidia emphasized that demand for its AI systems remains robust across various segments, including enterprises investing in AI infrastructure and “neoclouds” – companies challenging the established giants with specialized AI services. Furthermore, Nvidia is targeting foreign governments, anticipating $20 billion in revenue for “sovereign AI” initiatives this year. This broader demand profile is a key component of Nvidia’s ongoing revenue growth, according to CFO Kress.
CEO Jensen Huang projected a staggering $3 to $4 trillion market for AI infrastructure by the decade’s end. He estimates Nvidia could capture approximately 70% of the total cost of a $50 billion AI-focused data center, not only through its GPUs but also through the array of other chips it supplies.
Huang justified this ambitious forecast based on the substantial capital expenditure commitments from hyperscalers, estimated at $600 billion this year alone, and the increasing participation of new customer segments, such as enterprises and international cloud providers.
“As you know, the capex of just the top four hyperscalers has doubled in two years as the AI revolution went into full steam,” Huang stated, underscoring the profound impact of AI on infrastructure investments.
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/8217.html
