Nvidia’s earnings validate strong AI demand, but financing and concentration still require scrutiny. (Photo by Justin Sullivan/Getty Images)
Getty Images
In the days leading to Nvidia’s second-quarter earnings announcement, the market yearned for reassurance that AI demand remained strong. The results came in. Demand for semiconductor chips far exceeded supply, sending a strong message to anxious investors. Breaking with tradition, the company’s CEO, Jensen Huang, who previously refrained from commenting about future demand, provided strong forward guidance for data-center revenue, predicting 70% growth in the 2028 fiscal year (which oddly starts in late January 2027). This is nothing short of spectacular for a company of its size.
This confirms that the AI boom is not short of demand, at least in the next several months. The company reported $96.2 billion in revenue for the quarter, an increase of 106% year over year, of which $89 billion came from its data-center business. Nvidia’s performance calmed near-term concerns about a demand deceleration, but signs of fragility still exist.
Before the earnings release, I proposed a scorecard covering four tests: data-center demand vitality, quality of execution during the product platform transition to Vera Rubin, the degree of customer concentration and the nature of the ecosystem financing.
Data Center Demand
Product demand accelerated in Nvidia’s two largest customer segments. Hyperscale customers, according to the company’s quarterly 10-Q filing, purchased $48.7 billion, up 13% since last quarter. AI clouds, industrial customers and enterprises (ACIE) revenue increased 25% quarter-to-quarter and 138% since last year, totaling $40.3 billion. The faster growth in ACIE purchases confirms that acceleration is increasing outside the hyperscale customers. These are the large cloud providers that support the AI frontier labs such as Anthropic and OpenAI. Together, hyperscale customers and ACIE make up the data center client segment.
During the earnings call, chief financial officer Colette Kress called the 70% revenue growth forecast for fiscal 2028 “a supply-constrained outlook.” Huang confirmed in an interview with CNBC’s Jim Cramer that demand is “super strong” and accelerating. The main criticism in an otherwise stellar financial report was that profit margins got worse and are expected to fall from 75% to 74% in 3Q, the current quarter, and then stabilize between 71% and 72% in 4Q. This gross margin erosion is driven primarily by memory shortages driving higher costs.
Nvidia Product Transition And Execution Risk
Vera Rubin is Nvidia’s latest platform, succeeding the Blackwell generation. Rubin reached full production in August 2026, and already has operational systems at clients like CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius. The new line is expected to generate 20% of data center revenue this quarter in what Kress called “the fastest product ramp in Nvidia’s history.”
Buyers are not waiting on the sidelines for the next architecture and the product transition to Rubin is not freezing Blackwell purchases. Operational execution is now the key to success, as the company translates its impressive sales forecast into shipments. The supply chain is stretched to its limit as Nvidia seeks to satisfy customer deployments while managing margins.
Nvidia Customer Concentration Creating Financial Risk
A small number of large clients drives the bulk of Nvidia’s business. One customer accounted for 16% of Nvidia’s revenue last quarter, and five customers made up 70% of its accounts receivable, according to the 10-Q report.
Days sales outstanding rose to 60 days as Nvidia stretched payment terms for select investment-grade clients to up to one year. A more generous credit policy does not necessarily indicate poor revenue quality or financial distress, as healthy vendors tend to leverage their balance sheets, extending credit to customers or accepting slower collections in exchange for sales volumes. Vendor credit helps finance clients’ purchases but moves risk to Nvidia’s balance sheet. In this case, the size and concentration are the issues. Next quarter, investors need to watch whether Nvidia will continue to grow its customer base beyond a handful of large clients.
Customer and Partner Financing By Nvidia
Nvidia funds a large part of the AI infrastructure buildout. Its 10-Q shows that the company helps clients secure data center capacity, land and power with $36 billion of “AI-cloud commitments,” an item whose performance Nvidia cannot control. This support to clients adds to $366 billion in “future commitments,” most of it from $279 billion for supply and capacity.
The company extends customer guarantees with a maximum exposure of $108.5 billion. These guarantees are dominated by $105 billion associated with one 4.25-gigawatt OpenAI project in Ohio.
This level of direct exposure has raised concerns in the analyst community. To diversify exposure, Nvidia recently signed agreements to create an alternative source of capital for its customers and raise more than $500 billion, partnering with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR. The company’s CFO addressed concerns about “circular financing” in her presentation, stressing that third-party capital providers will do their independent underwriting and that Nvidia’s compute is durable and fungible.
The Nvidia Scorecard
On the positive side, data center demand is strong and growing beyond hyperscale customers. Additionally, the product transition is being lifted by solid initial sales and has not experienced technical issues. The weak links are the persistent customer and receivables concentration and the high level of direct exposure as Nvidia finances the AI infrastructure boom.
Leave a Reply