AI Server Sector Reaches Profit Realization Phase, Lenovo Faces Crucial Test

Deep News07-13

The AI infrastructure supply chain in China has delivered two highly impactful performance reports over the past week.

Inspur Information forecasts its net profit attributable to shareholders for the first half of 2026 to be between RMB 2.6 billion and RMB 3.1 billion, representing a year-on-year increase of 226% to 288%. The company attributes this growth to favorable industry conditions, enhanced product value, and improved supply chain capabilities.

Subsequently, Industrial Fulian expects its first-half net profit to reach between RMB 23.4 billion and RMB 24.4 billion, a surge of 93% to 101%, marking the first time it has surpassed the RMB 200 billion mark. More indicative of industry demand strength is that Industrial Fulian's AI server revenue from cloud service providers grew over 230% year-on-year in the first half, while shipments of data center switches with speeds above 800G increased by 1.4 times.

These figures are shifting a long-held market perception of the server industry: it is no longer just a hardware assembly business characterized by fast revenue growth but low profit margins. At least in the financial statements of leading companies, AI server demand is now translating into simultaneous revenue growth, product mix improvement, and profit elasticity.

This has also raised investor expectations for Lenovo Group's August earnings report. For Lenovo, the strong performance of its peers serves both as a positive signal and a higher benchmark for its own results. The market now expects not just continued server revenue growth, but proof that scale expansion is consistently translating into profit and cash flow. The question is whether the profit inflection point seen last quarter was a temporary improvement or the beginning of a sustainable shift in profit structure.

Industry Momentum Moves Beyond Order Stories

Recently, the market has been influenced by a narrative of "computing power surplus." This stemmed from Meta's plan to sell excess AI computing capacity to external clients, quickly sparking investor concerns about an oversupply of computing power and a capital expenditure bubble, leading to significant volatility in related AI infrastructure, chip, and memory stocks. For a sector that has seen sustained gains, "excess capacity" is easily interpreted as demand nearing its peak.

However, the subsequent earnings reports from the supply chain present a different signal. The performance forecasts from Inspur Information and Industrial Fulian show that AI server demand has moved from orders to the income statement. Meanwhile, Samsung Electronics expects a record-breaking second-quarter operating profit of 89.4 trillion won, driven in part by AI-fueled server memory demand and rising storage prices. The simultaneous strong profit surges from server OEMs, system manufacturers, and upstream memory providers suggest no clear evidence yet that the entire AI hardware demand has peaked.

Concurrently, the global server market is showing an unusual divergence in data: limited growth in unit shipments but rapid growth in market value.

IDC data shows global server market revenue reached $122.6 billion in Q1 2026, up 30.4% year-on-year, while shipments grew only 3.3%. Accelerated computing servers now contribute over 70% of market revenue. The revenue growth rate far outpacing shipment growth indicates this industry expansion is not driven by a massive increase in traditional server volumes, but by rising per-unit value propelled by GPUs, networking, storage, liquid cooling, and power systems.

This explains why the profit growth rates for Industrial Fulian and Inspur Information significantly exceed the typical levels for traditional hardware manufacturing.

As servers evolve from standardized devices into complex systems integrating computing, networking, storage, cooling, and software, suppliers offer more than just assembly capacity. Companies capable of participating in product design, securing key components, achieving rack-level integration, and executing large-scale deliveries across markets are better positioned to capture higher product value and operating leverage.

In a July Asia Technology Strategy report, Goldman Sachs views the AI server and data center hardware cycle as still "large and long." The firm has not yet observed the two typical signals of a cycle peak: component supply beginning to outstrip demand, and a shift in technology focus from pursuing higher performance to simply reducing costs. Conversely, supply tightness is spreading from core areas like memory to optical communications, substrates, electronic materials, power supplies, and other peripheral components.

Demand also shows no signs of significant contraction. Goldman Sachs estimates that combined capital expenditure from major US cloud service providers will grow about 76% in 2026, up from 74% in 2025. Demand for AI server racks could rise from approximately 55,000 units in 2026 to 105,000 in 2027 and 163,000 in 2028.

This data cannot rule out localized overinvestment or prove that all AI asset valuations are justified. However, from the perspective of cloud provider spending, server market revenue, and key component supply and demand, the industry currently appears closer to one that is still expanding but beginning to be tested on capital returns, rather than a market where demand has already peaked.

These industry signals provide server manufacturers with high revenue visibility, but do not guarantee equal profit growth for every company.

AI servers have high core component costs and high customer concentration. Large cloud service providers possess strong bargaining power, meaning server companies might secure billions in revenue but retain limited profit. Therefore, at this stage, the focus is shifting from who secures the largest orders to who can preserve more profit amidst rapid technological upgrades, large-scale delivery, and supply constraints.

The forecasts from Inspur and Industrial Fulian essentially set a higher benchmark for Lenovo. If industry demand has not peaked and some leading players are already demonstrating significant operating leverage, Lenovo's infrastructure business also needs to prove that its growth is not merely about exchanging more working capital for more revenue.

Lenovo's Server Business Awaits Performance Validation

Lenovo achieved a crucial turnaround to profitability for its server business in the last fiscal year. For the fiscal year ending March 31, 2026, Lenovo's revenue grew 20% to $83.1 billion, with adjusted net profit up 42% to $2.0 billion—profit growth roughly double revenue growth. The Infrastructure Solutions Group (ISG) reported full-year revenue of $19.2 billion, a 32% increase, achieving an operating profit of $73 million; Q4 operating profit further rose to a record $202 million. The company also holds over $140 billion in AI server backlog orders.

However, for the full year, ISG's operating margin remained at only about 0.4%. This indicates the profit trajectory has changed, but the profit base remains thin. A single product delay, cost increase, or change in order mix could significantly impact margins.

Therefore, the core of the August Q1 report is not whether ISG revenue maintains double-digit growth, but whether the margin level established last quarter can be sustained.

CLSA offers a relatively positive forecast on this issue. The institution believes that demand for GB200 and GB300 servers from new US cloud customers could drive Lenovo's Q1 server business revenue to grow 45% year-on-year, higher than the previous 35% estimate. The operating margin is forecast to be revised up from an earlier estimate of 3.1% to 5.2%, corresponding to an operating profit increase from $180 million to $320 million. CLSA also expects Lenovo's FY27 server business revenue to grow 45% to $28 billion, with AI servers contributing over 30%, accounting for about 10% of group revenue.

If this prediction is validated, its significance would extend beyond a single quarter of earnings beat. A 45% revenue growth primarily proves strong demand; achieving a 5.2% operating margin signifies a change in the quality of incremental revenue. It could reflect a higher proportion of more valuable AI servers or indicate that Lenovo is gaining scale advantages in procurement, R&D, manufacturing, and delivery.

Based on this forecast, CLSA raised its target price for Lenovo Group from HK$30 to HK$36, maintaining an "Outperform" rating and upgrading it to a top pick in the China tech sector.

In retrospect, this is the most important market implication of the Inspur and Industrial Fulian forecasts for Lenovo. They cannot directly predict Lenovo's profit due to differences in clients, products, and accounting methods among the three companies. However, they prove that the server industry currently possesses the conditions for product mix improvement and profit release. Whether Lenovo can achieve similar operating leverage will determine if the market continues to view it primarily as a cyclical PC hardware company.

Two issues still require attention. First, rapid growth may primarily come from a few large cloud customers. Such orders are large in scale but often come with lower gross margins and longer payment cycles. Investors need to observe whether accounts receivable, inventory, and operating cash flow improve in sync with profits.

Second, supply constraints represent both a competitive moat and a cost pressure. TrendForce estimates Q2 2026 traditional DRAM contract prices rose 58% to 63% quarter-on-quarter, with NAND Flash up 70% to 75%. Goldman Sachs believes memory supply tightness may persist into 2027 or even 2028, with clients increasingly locking in supply through long-term agreements.

Lenovo's sensitivity to memory prices is particularly complex. AI servers require large amounts of DRAM, SSDs, and high-bandwidth memory, while the PC business also relies heavily on memory and flash storage. In other words, rising memory prices could, on one hand, benefit leading players with stable supply and procurement scale to expand market share, but on the other hand, simultaneously pressure margins in both Lenovo's server and PC businesses.

Global PC shipments in Q2 already declined 4.9% year-on-year to 68.2 million units. While Lenovo's shipments fell 2.1%, its outperformance relative to the industry increased its market share from 23.7% to 24.4%. This reflects its share resilience but also means the PC business cannot fully escape the impact of declining industry volumes and rising costs.

Therefore, the Intelligent Devices Group (IDG) remains the other crucial half of this earnings report. Whether server profit improvement translates into accelerated group profit growth largely depends on whether the PC business can absorb rising memory costs through product price increases, a higher mix of premium products, and procurement advantages.

From Server Profitability to Group Profit Structure Reshaping

For Lenovo, the significance of the August 13th earnings report extends beyond whether ISG can deliver another profitable quarter; it lies in whether server business improvement can become the starting point for a sustained rise in group profitability.

At its June Investor Day, Lenovo management outlined a clearer long-term financial path: achieving $100 billion in revenue with a net profit margin exceeding 3% within one to two years; reaching $130 billion in revenue with a net margin over 5% within three to five years; and ultimately exceeding $150 billion in revenue with a net margin above 8% in the longer term. Compared to last fiscal year's approximately $83.1 billion revenue and about 2.5% adjusted net margin, this signifies Lenovo's goal is not just to expand scale but to significantly increase the profit generated per dollar of revenue.

This path can be distilled into a business portfolio logic: IDG maintains market share and the profit base; ISG transitions from a revenue engine to a profit engine; and the Solutions and Services Group (SSG) elevates overall group profitability through high-margin, recurring revenue. As the proportion of non-PC business increases, expanding revenue scale will further dilute R&D, sales, and administrative expenses.

ISG is the most critical variable. The PC business is relatively mature, and the service business has stable profitability; only when the server business shifts from breakeven to sustained profitability can the group's net profit margin undergo a structural uplift. This is why the market is particularly focused on ISG's margin this quarter: it affects not just one quarter's profit but relates to the realistic foundation of management's long-term 5% net profit margin target.

Therefore, the August report will serve as the first practical test following the Investor Day. The market needs to observe not just whether ISG remains profitable, but also whether IDG can digest memory price pressures, whether SSG can maintain growth faster than the group, and whether operating cash flow improves in line with profits.

A single earnings report cannot prove long-term targets are achieved, but it can provide important clues: whether Lenovo's current profit improvement is merely a cyclical boost from AI server momentum or marks the beginning of a fundamental reshaping of the group's profit structure.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment