Nvidia says two unnamed customers drove nearly 40% of Q2 revenue

Nvidia posts record-breaking second quarter

Nvidia reported an astonishing $46.7 billion in revenue for the quarter ending July 27, a 56% increase compared to last year. Net income also surged 59% year over year to $26.4 billion.

This remarkable growth was powered by exploding demand for artificial intelligence data centers, where Nvidia’s GPUs have become essential.

Despite headwinds in China due to U.S. export controls, Nvidia continues to break records, showing how critical its technology has become worldwide.

Two customers account for almost 40 percent of sales

The headline detail from Nvidia’s SEC filing is striking: just two unnamed customers together represented 39% of total revenue in Q2. One buyer, identified only as “Customer A,” accounted for 23%, while “Customer B” comprised 16%.

That means nearly $18.2 billion flowed from just two sources. The disclosure highlights Nvidia’s enormous dependence on a handful of big buyers, even as overall sales are soaring at unprecedented levels across industries.

Concentration has grown since last year

What makes this disclosure even more noteworthy is how much those customer shares have increased. In the same quarter a year earlier, the top two customers comprised only 14% and 11% of sales.

That’s a jump of nearly 14 percentage points year over year. Nvidia has always had periods of concentrated customer spending, but the pace at which these buyers are scaling their purchases is drawing fresh scrutiny from analysts and regulators alike.

Customers are listed as direct buyers

Nvidia described both Customer A and Customer B as “direct customers.” That means they buy chips straight from Nvidia, unlike “indirect customers” such as Microsoft, Google, or Amazon, who often source GPUs through distributors or system integrators.

Direct customers can include OEMs, ODMs, add-in board makers, or large distributors. This complicates speculation about who exactly these mystery clients are, though their enormous spending suggests they serve hyperscale cloud or AI research organizations.

Indirect cloud giants may still drive the spending

Although Nvidia clarified that Customers A and B are direct purchasers, that doesn’t mean cloud providers aren’t behind the massive orders.

These chips could ultimately be destined for Amazon AWS, Microsoft Azure, or Google Cloud data centers.

Nvidia acknowledged that “large cloud service providers” represented about 50% of its data center revenue in Q2, making up 88% of overall revenue. So while customers are labeled “direct,” the real demand may come from hyperscale AI buildouts.

Four other customers make up another large slice

Beyond the two top buyers, Nvidia’s SEC filing shows four other customers contributed double-digit revenue shares in Q2: 14%, 11%, 11%, and 10% respectively.

Altogether, just six customers accounted for more than 80% of Nvidia’s quarterly revenue. That’s an extraordinary concentration level for a company valued at over $3 trillion.

While demand for AI chips is global, Nvidia’s results hinge heavily on spending decisions made by fewer than a dozen organizations.

Analysts warn of concentration risk

Industry analysts quickly pointed out that depending on so few customers presents a double-edged sword. Dave Novosel of Gimme Credit told Fortune that “concentration of revenue among such a small group of customers does present a significant risk.”

If one or two buyers were to cut back, Nvidia could see billions wiped from its top line. However, he also noted that these buyers are among the most cash-rich companies in the world, mitigating immediate danger.

Big buyers are flush with cash

The good news for Nvidia is that its largest customers are not struggling for funds. These companies have vast war chests and generate billions in free cash flow each quarter.

With cloud capital expenditure budgets exploding, many analysts expect them to keep spending aggressively on AI infrastructure for years.

This creates a paradox for Nvidia: while the dependence on a handful of buyers looks risky, those buyers are arguably the most reliable spenders possible.

Cloud spending fuels the AI gold rush

Microsoft, Amazon, Google, and Oracle are racing to expand their AI capacity. According to Nvidia CEO Jensen Huang, the top four hyperscalers have doubled their capital spending in just two years.

Much of this goes straight into GPU-driven data centers. While Nvidia has not confirmed if these cloud leaders are behind Customers A and B, the timing and sheer scale of orders strongly suggest they play a decisive role in fueling Nvidia’s revenue surge.

Data center business dominates Nvidia revenue

The center of Nvidia’s empire is its data center division, which brought in $41.1 billion in Q2, up 5% from Q1 and up 56% from a year ago. Just two years ago, that figure was only $10.3 billion.

The leap reflects how rapidly AI has transformed the company from a gaming-focused GPU maker to the backbone of global computing infrastructure.

Cloud providers, enterprises, and governments rely on Nvidia chips to run large language models and AI systems.

Nvidia maintains a near monopoly on AI GPUs

Nvidia currently controls more than 90% of the market for AI GPUs. Rivals like AMD and Intel have products in development, while Google and Amazon are designing custom chips for their data centers.

But for now, Nvidia’s CUDA software ecosystem and hardware performance create a moat that competitors struggle to breach.

This dominance explains why customers are willing to concentrate so much spending on Nvidia, despite the risks of over-reliance on one supplier.

China remains a missing piece of the puzzle

One remarkable detail is that Nvidia achieved record sales without counting on China. Due to U.S. export restrictions, it has been unable to ship its most advanced H20 chips there, leading to a $4.5 billion write-down in inventory.

Nvidia estimates the Chinese AI market could be worth $50 billion annually, growing at 50% annually. If U.S. regulators eventually allow broader sales, Nvidia could add billions more to its staggering growth.

Blackwell chips drive the latest surge

Nvidia credited much of its success to the rollout of its Blackwell architecture GPUs, which set benchmarks for training and inference performance in AI.

These chips are powerful and optimized for efficiency, which appeals to hyperscalers scaling massive clusters.

Blackwell systems can take up as much as 70% of the total cost of a $50 billion AI data center. This puts Nvidia firmly at the heart of global AI expansion.

Networking products shine alongside GPUs

While GPUs get the headlines, Nvidia’s networking portfolio is quietly booming. Data center networking revenue nearly doubled to $7.3 billion in Q2.

Products like InfiniBand, NVLink, and Spectrum-X Ethernet are critical for connecting massive GPU clusters. By controlling both computing and interconnecting, Nvidia locks customers into its ecosystem.

Analysts note networking could soon rival GPUs as a key growth engine, giving Nvidia multiple profit streams from the same AI infrastructure buildouts.

Gaming and other businesses show strong growth

Although AI dominates, Nvidia’s traditional segments are still expanding. Gaming revenue jumped 49% yearly to $4.3 billion, while professional visualization grew 32%.

Automotive sales rose 69% to $586 million, fueled by self-driving solutions and partnerships with carmakers. These areas highlight Nvidia’s versatility, even if they remain small compared to data centers.

By maintaining growth across multiple verticals, Nvidia reinforces its image as a diversified technology leader rather than a one-trick AI company.

See why Nvidia’s CEO believes AI will create millionaires even faster than the internet once did.

The bottom line is a massive opportunity with some risk

Nvidia’s Q2 results highlight both its strengths and vulnerabilities. Record revenue, dominance in AI GPUs, and booming demand make it the world’s most crucial chipmaker.

However, nearly 40% of sales come from just two customers, raising fundamental questions about concentration risk. Those customers are spending heavily and show no signs of slowing.

As long as the AI boom continues, Nvidia looks set to keep rewriting records, though its reliance on a few buyers bears watching.

Learn why Nvidia excludes China from revenue outlook as tensions reshape strategy.

What do you think about Nvidia getting customers, which is causing massive profits for the company? Please share your thoughts and drop a comment.

Read More From This Brand:

• Trump’s support for Nvidia and AMD adds pressure to global chip rivals
• Chinese tech firms told to boycott Nvidia under new government order
• Trump signals support for Nvidia’s limited Blackwell sales to China

Don’t forget to follow us for more exclusive content on MSN.

If you liked this story, you’ll LOVE our FREE emails. Join today and be the first to get stories like this one.

This slideshow was made with AI assistance and human editing.