AI server shipment surge, why Dell's unexpectedly "slip" in profit expectations | AI Dehydration

Author: Zhang Yifan

Editor: Shen Siqi

Source: Hard AI

Recently, the three major AI PC manufacturers, Lenovo, Dell, and HP, have successively disclosed their latest quarterly reports.

• Lenovo's performance showed that driven by AI, the group's revenue in this quarter increased by 9.5%, net profit doubled, and it forecasted potential orders of $7 billion for AI servers. After the meeting, the stock price rose by 9%.
• Similarly, HP, benefiting from AI, reported a net profit per share of $0.82, higher than the same period last year. The CEO stated that in the fiscal year 2024, up to 10% of HP's sales will come from AIPC, leading to a 12% increase in stock price to a new high.

Among the three PC giants, Dell, whose AI servers exceeded market expectations, turned out to be the one that fell short of expectations. This quarter, poor AI performance led to a decrease in profit, causing a 20% drop in after-hours trading.

1. Profit Falls Short of Expectations, Dell Plunges 20% After Hours

Dell's latest quarterly earnings meeting showed that both storage revenue and operating profit were below market expectations.

1) Storage Revenue Falls Short of Expectations

Storage remained flat compared to the previous quarter, with no growth, far below the market's expected 18% growth.

The company previously revealed at a conference —

"For every $1 of AI server, the company will bring in $2 of services, storage, and other higher-margin businesses."

Considering that AI server shipments this quarter increased by $1.7 billion compared to the same period last year, it was speculated based on the company's previous statement that storage revenue associated with AI server shipments should have seen a significant year-on-year increase.

However, actual operational data shows that storage revenue did not change, falling far below the market's pre-meeting expectation of 18% growth.

In addition, the AI server supply chain shows that HBM and DDR memory are in short supply due to AI catalysis, yet Dell did not see a significant change in revenue amidst this wave of storage shipments.

2) Profit Falls Short of Expectations

Although AI server shipments increased by $1.7 billion compared to the same period last year, profits decreased instead of increasing, well below the market's expected 17% growth.

Analysts raised questions at the earnings meeting —

According to the data disclosed by the company, the AI server business grew from $0 to $1.7 billion, while traditional servers and storage servers remained flat compared to the previous quarter.

This means that the real change in the entire ISG business this quarter was the addition of $1.7 billion from AI servers However, despite an increase in revenue of $1.7 billion, operating profit remained flat compared to the previous quarter. Analysts questioned whether this meant the operating profit margin for AI servers was actually 0.

While the company responded that "large customers and big orders led to lower profit margins," the market does not seem to agree, judging by the 20% drop in after-hours trading.

II. Micron's New DRAM Expansion Plan

According to Japanese media reports, Micron plans to build a new DRAM chip production factory in Hiroshima Prefecture, Japan, with the goal of starting operations by the end of 2027.

Following the validation of Micron's HBM3E memory by NVIDIA, the market is closely watching Micron's expansion plans.

Goldman Sachs predicts that by 2024, Micron's HBM capacity will increase by 203% year-on-year, with capacity growth continuing until 2026.

However, at this month's earnings call, Micron did not disclose any DRAM-related capacity expansion plans, only mentioning that they would consider market supply and demand conditions to plan capacity according to demand, to prevent an oversupply situation in the market.

Perhaps due to the intense demand for HBM.

Recently, Japanese media revealed that Micron plans to build a DRAM factory in Japan, with the total investment estimated to be between 600 billion and 800 billion Japanese yen (approximately $5.1 billion). The construction plan is set to start in early 2026 and will be equipped with extreme ultraviolet lithography (EUV) equipment.

In addition, the Japanese government has approved subsidies of up to 192 billion Japanese yen (approximately $1.3 billion) to support Micron's production of next-generation chips at the Hiroshima factory.

The Japanese factory is another new addition to Micron's factories in the United States and India in recent years.

III. AI Empowering Pathological Detection

AI's analytical advantages in data, images, videos, and other fields can assist doctors in diagnosing diseases more accurately and quickly.

Recently, Goldman Sachs released a report on "AI + Pathological Detection," predicting that by 2030, the total addressable market for digital pathology will reach around $1 billion, and by 2035, it will reach around $3 billion.

The report mentions that in the past, pathological analysis relied on pathologists examining tissue samples under a microscope. However, due to a shortage of pathologists and the increasing volume of data from various sources such as digital images and patient records, this method is facing challenges Digital pathology can convert images on glass slides into digital images, which is achieved by creating "whole slide imaging" or WSI using a scanner.

The report emphasizes the potential of AI in improving the efficiency and accuracy of pathology workflows. For example, a study found that compared to human pathologists, AI algorithms have a much higher accuracy in detecting metastases in lymph node samples (92% versus 73% sensitivity).

In addition, Paige, a company focusing on "AI + healthcare", has developed an FDA-approved artificial intelligence system specifically for prostate cancer. The system has also achieved remarkable results: 99.3% sensitivity, a 70% reduction in cancer detection errors, and a 65.5% reduction in diagnostic time.

Goldman Sachs pointed out that companies such as Guardant Health (GH), Veracyte (VCYT), and NeoGenomics (NEO) may benefit from this growth as they have advantages in WSI digital repositories, AI analytics technology, and biopharmaceutical services