Goldman Sachs Downgrades Inspur to Neutral: AI Chip Transition Pressures Margins, Valuation Already Reflects Growth

Overview

This report primarily analyzes Inspur's market position, product structure adjustments, and financial outlook in the AI server sector. Goldman Sachs recognizes Inspur's status as a leader in China's AI server market, particularly its advantages in inference servers and open compute architecture. However, considering the transition of AI servers from global top-tier GPUs to domestic AI chips, which will lead to a decline in Average Selling Price (ASP), and the rising proportion of direct sales to Chinese Cloud Service Providers (CSPs) which erodes high-margin channel business, resulting in long-term pressure on blended gross margins. Based on this, Goldman Sachs has lowered earnings forecasts and the target price, downgrading the rating from 'Buy' to 'Neutral', believing the current stock price fully reflects future growth expectations.

Core views

Core View 1: Chip transition and changing customer demands doubly suppress profit margins. Inspur's AI servers are undergoing a platform transition from global top-tier GPUs like NVIDIA to domestic AI chips. While domestic chips have lower costs, this has also led to a decline in the overall ASP of servers. Furthermore, the increasing share of AI server procurement by Chinese Cloud Service Providers (CSPs), which mostly follows a direct sales model lacking the brand premium of traditional channel business, is dragging down the company's blended gross margin. Goldman Sachs expects the company's gross margin to fall from the 10-15% range in 2015-2023 to 7%/5% in 2024/25, and remain at a low level of 5-6% in 2026-28. Core View 2: Continuous optimization of product structure with rising share of high-priced servers. Despite margin pressure, Inspur's product mix is improving significantly. Data shows that in 2025, shipments of inference AI servers priced between $50k-$100k grew 31% YoY, and shipments of training AI servers priced between $100k-$250k grew 47% YoY. The contribution of low-priced servers (<$50k) dropped from 98% in 2023 to 94% in 2025. Goldman Sachs expects shipments of servers priced above $50k to grow at a 55% CAGR in 2026-28, with their revenue contribution rising from 17% in 2025 to 39% in 2028. Core View 3: Earnings forecast lowered, valuation returns to rationality. Based on the aforementioned gross margin and ASP pressures, Goldman Sachs has cut Inspur's 2026/27 net profit forecasts by 20% and 2% respectively. However, thanks to improved operational efficiency (operating expense ratio dropped to 4.0% in Q1 2026), the 2028 net profit forecast was slightly raised by 4%. Regarding valuation, Goldman Sachs uses a method linking peer P/E ratios to forward net profit growth, lowering the target P/E multiple from 26.4x to 23.8x (based on 2027 expected earnings), deriving a new 12-month target price of 76.5 CNY. Given the limited upside implied by the current stock price and the fact that geopolitical supply chain uncertainties have not been fully eliminated, the firm believes positives are largely priced in.

Analysis framework

Goldman Sachs' analysis logic follows a path combining 'volume-price breakdown' and 'peer relative valuation'. First, by breaking down AI server price bands (e.g., <$50k, $50-100k, >$100k), it quantifies shipment growth and ASP changes across different price segments to derive revenue structure and gross margin trends. Second, it deeply analyzes the specific impact of customer structure changes (CSP direct sales vs. channel distribution) on blended gross margins. Finally, at the valuation level, instead of relying solely on historical averages, it establishes a correlation model between peer companies' 'Price-to-Earnings (PE)' and 'Forward Net Profit Growth (NI YoY)'. Based on Inspur's expected 36% net profit growth, it dynamically determines a reasonable target PE multiple to judge whether the current stock price is overvalued or undervalued.

Methodology notes

• Industry/Sector Analysis FrameworkVolume-Price Breakdown

  Volume-Price Breakdown

  Decomposes revenue growth into two dimensions: 'volume growth' and 'Average Selling Price (ASP) change'. In this report, it is used to analyze the specific impact of declining ASP on total revenue and gross profit during the transition of AI servers to domestic chips, even if volume may grow.

• Valuation MethodPE/PEG valuation

  PE/PEG Valuation

  Determines reasonable valuation multiples by comparing the relationship between Price-to-Earnings (PE) ratios and Net Profit Growth rates of companies in the same industry. The report utilizes the correlation between peer PE and forward net profit growth to set a target PE of 23.8x for Inspur, reflecting the logic of matching growth with valuation.

• Competition and Strategy FrameworkValue chain analysis

  Value Chain Analysis

  Analyzes a company's position in the industrial chain and profit distribution. The report points out that Inspur's shift from channel-focused branded business to CSP-focused direct sales business, while closer to major clients in the value chain, has lost channel brand premium capabilities, leading to a structural decline in gross margins.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Inspur (000977.SS)

  Interwoven benefit/damage logic: Benefits from growth in China's AI infrastructure spending and explosion in inference demand, but suffers from ASP decline due to chip transition and margin compression from rising CSP direct sales share.

  Strengths

  Leading market share in China's AI server market; Open Compute Architecture wins international awards, strong R&D capability, adaptable to multi-chip platforms; Rapid rise in share of high-priced servers.

  Weaknesses

  Structural decline in blended gross margin to 5-6%; Increased dependence on a single region (China) and specific customer group (CSP); Supply chain uncertainties brought by geopolitics.

  Comparison

  Compared to peers, Inspur has advantages in scale and localization adaptability, but its gross margin level is lower than during the period with stronger channel model brand premiums.

  Risks

  Generative AI demand falls short of expectations; Chip transition speed is slower than expected; Intensified competition in domestic AI server market.

Key data

• 2026/27E Net Profit Forecast Adjustment-20% / -2%Reflects ASP decline and gross margin pressure
• 2026-28E Expected Blended Gross Margin5-6%Significant decline from 10-15% in 2015-2023
• New Target Price76.5 CNYPreviously 86.5 CNY, implies 9.7% upside
• Target P/E Ratio (2027E)23.8xPreviously implied 2027E PE of 26.4x
• Revenue Share Forecast for >$50k Servers39% (2028E)17% in 2025, reflecting product mix premiumization
• Shipment CAGR for >$50k Servers+55%Expected CAGR for 2026-28

Impact & implications

For Inspur, the short term involves a painful transition period of margin contraction, but in the long run, its R&D advantages in open compute architecture will help adapt to China's diversified AI chip ecosystem. For investors, the report highlights the risk of overall valuation de-rating in the AI server sector, as high growth may not sustain high gross margins amid geopolitical tensions and supply chain transitions. At current valuation levels, the stock price has fairly reflected the expected ~36% annual net profit growth over the next two years; further upside requires better-than-expected product mix upgrades or a smoother chip transition.

Risks

• China's generative AI demand is stronger/weaker than expected, affecting CSP procurement of AI servers and company revenue estimates.
• The speed of AI server transition from global top-tier GPUs to domestic chips is faster/slower than expected, affecting revenue and gross margin estimates.
• The intensity of competition in China's AI server market is higher/lower than expected, affecting shipments, market share, ASP, and gross margins.

What to watch

• Whether the company can achieve a smoother AI server model transition than expected.
• Whether the pace of product mix upgrades in server or storage businesses is faster than expected.
• Return on Invested Capital (ROIC) and application effectiveness of Chinese Cloud Service Providers' investments in generative AI.