The AI server business is booming, but all the money is being made by NVIDIA

Against the backdrop of surging demand for artificial intelligence, major global AI server manufacturers are facing a common challenge: despite significant revenue growth, profit margins are continuously narrowing.

As Hewlett Packard Enterprise (HPE), Dell, and AMD release their financial reports, a harsh reality emerges: despite a surge in AI server orders, the high costs of NVIDIA chips and intense market competition are making hardware manufacturers' profit margins exceptionally thin.

Hewlett Packard Enterprise's third-quarter financial report released on Wednesday showed that the company's revenue increased by 18% year-on-year to $9.14 billion, with earnings per share of $0.44, both exceeding analysts' expectations. However, its server division's operating profit margin fell from 10.8% in the same period last year to 6.4%.

In response to investors' concerns, HPE CEO Antonio Neri urgently reassured the market during an analyst conference call, promising that profit margins would return to around 10% by the end of this fiscal quarter. This statement temporarily stabilized the stock price, with HPE's shares rising 1.5% in after-hours trading.

This trend is spreading throughout the industry, with the fundamental reason being that the core of AI servers—high-performance GPU chips—is almost monopolized by NVIDIA. These expensive chips account for a large portion of server costs, leaving the vast majority of value upstream in the supply chain.

The Dilemma of "Increased Revenue Without Increased Profit"

For original equipment manufacturers (OEMs) of servers, the current AI market presents an awkward situation of "increased revenue without increased profit."

In addition to HPE, whose profit margins have nearly "halved," another server giant, AMD, is facing the same challenge. Although its revenue surged by 46.59% year-on-year in the fourth quarter of 2025, its gross margin has fallen to 9.7%.

Dell's situation is similar, with its gross margin in the second quarter of 2026 dropping from 22% in the same period last year to 18.7%, which the company attributes to pricing pressure from AI servers.

This phenomenon of "high revenue, low profit" is becoming a common dilemma for AI hardware manufacturers, contrasting sharply with the astonishing profitability of chip giant NVIDIA.

NVIDIA, with a commanding 98% share of the data center GPU market, holds absolute pricing power. Its second-quarter 2026 financial report showed a non-GAAP gross margin of 72.7%, several times that of server manufacturers. Further data indicates that its latest Blackwell GPU platform can achieve a profit margin of 77.6% in AI inference workloads.

This stark disparity in profit distribution clearly illustrates the current state of the AI value chain:

NVIDIA: Gross margin exceeds 70%.

Dell: Gross margin approximately 18.7%.

HPE: Server division operating profit margin fluctuates between 6.4% and 10%

Super Micro Computer: Gross margin drops to 9.7%.

Behind the Profit Pressure

The profits of server manufacturers continue to be under pressure, mainly due to three structural factors.

First is the high component costs. The core of AI servers is NVIDIA's GPU, which is expensive and in short supply, leaving OEM manufacturers with almost no bargaining power. A report even pointed out that in the cloud service provider sector, hardware OEM manufacturers may lose $1 for every $7.9 of AI hardware revenue they earn, highlighting the asymmetry in cost structure.

Second is the intense market competition. To capture market share, server manufacturers have engaged in fierce price wars. The practice of offering significant discounts to secure large customer orders has further eroded already thin profits. For example, Dell's Infrastructure Solutions Group's operating profit margin has dropped to 8.8%.

Finally, there is the complexity of supply chain management. To meet the urgent delivery demands for AI components, manufacturers have to bear additional logistics costs, while challenges in inventory management have further increased operational costs and profit pressure.

For investors, the growth story of AI server manufacturers is becoming increasingly complex. Although the revenue forecasts for these companies are very optimistic, sustainability faces severe challenges. Analysts warn that maintaining profit margins at levels of 10-11% for a long time, like Super Micro Computer, will not be sufficient to support long-term technological innovation or provide substantial returns to shareholders.

In the grand scheme of AI, the role of hardware assemblers is more like "movers," while NVIDIA is the true "big winner."