Goldman Sachs lowers AI server shipment forecast, industry growth faces slowdown.
On March 24, Goldman Sachs' analyst team downgraded the sales forecast for rack-level AI servers, with expected shipments for 2025 and 2026 revised down from 31,000 units and 66,000 units to 19,000 units and 57,000 units (calculated based on 144-GPU equivalent). This adjustment is mainly due to the impact of the product transition period and uncertainties in supply and demand.
Goldman Sachs believes that while the second quarter of 2025 will be a strong quarter for Taiwan's ODM/thermal supply chain, it holds a more conservative outlook for annual shipments and expects the product transition period may again impact shipments in the third quarter of 2025.
As a result, Goldman Sachs has correspondingly lowered the target prices for Taiwan's ODM and thermal supply chain-related companies (Quanta, Hon Hai, FII, Wistron, Avic, and Shuanghong), with reductions ranging from 7% to 21%, and downgraded Quanta Computer's rating from &#34;Buy&#34; to &#34;Neutral.&#34; Additionally, Goldman Sachs believes that for investors, companies related to ASIC AI servers generally outperform GPU AI server suppliers, and this trend may continue.
Goldman Sachs pointed out:
<blockquote>
The demand for high-performance AI servers will not be completely replaced by rack-level forms, as some customers still prefer motherboard solutions for design flexibility. We remain optimistic about the prospects for AI inference and general servers, driven by the application of AI technology and the recovery of upgrade cycles.
</blockquote>
<h2>Goldman Sachs significantly lowers global AI server shipments</h2>
Goldman Sachs' analyst team expects that AI training servers will remain the main driver of market growth, primarily driven by the demand for computing power from continuously upgraded advanced AI models. However, the growth rate of shipments is expected to be lower than previously anticipated, mainly influenced by the following factors:
<blockquote>
<ol>
<li>Impact of product transition period: As the GPU platform transitions to the next generation of products in the second half of 2025, shipments may slow during the transition period;</li>
<li>Challenges of production complexity: The production complexity of full rack systems increases the uncertainty of capacity ramp-up;</li>
<li>Demand-side variables: With the release of more efficient AI models (such as DeepSeek), there is controversy over the market's demand for intensive computing power;</li>
<li>Tariff risks: ODM manufacturers need time to diversify production bases and improve yield rates, while cloud service providers and enterprise customers also need time to assess additional costs.</li>
</ol>
</blockquote>
Therefore, Goldman Sachs also adjusted the revenue for AI training servers, expecting related revenue to grow by 30% year-on-year in 2025, reaching $160 billion; and grow by 63% year-on-year in 2026, reaching $260 billion. Previous forecasts were $179 billion and $248 billion, respectively.
**
<blockquote>
<ul>
<li>Rack-level AI servers: The expected shipment volume for 2025-26 is 19,000/57,000 racks (calculated based on 144-GPU equivalent), with a market size of $54 billion/$156 billion (previously forecasted as $88 billion/$182 billion).</li>
<li>High-power AI servers: The expected shipment volume for 2025-26 is 423,000/423,000 units (calculated based on 8PU equivalent), with a market size of $106 billion/$104 billion (previously forecasted as $90 billion/$66 billion).</li>
<li>AI inference servers: The expected sales for 2025-26 will grow by 41%/39%, and the value will grow by 105%/33%, thanks to the continuous expansion of application areas.</li>
<li>General servers: The expected sales for 2025-26 will grow by 6%/4%, and revenue will increase by 9%/7% year-on-year, benefiting from the gradual recovery of replacement cycles and the launch of new CPU platforms.</li>
</ul>
</blockquote>
<h2>Overall Downgrade of Supply Chain Stock Price Expectations</h2>
Based on the above adjustments to the AI server market outlook, Goldman Sachs has downgraded the earnings forecasts and target prices of several Taiwanese AI server supply chain companies, with Quanta's rating downgraded to neutral.
Goldman Sachs believes that in the overall coverage of the Greater China technology sector, Quanta's target price has relatively limited upside potential. Goldman Sachs expects uncertainty in the demand for rack-level AI servers and that the upcoming production transition may lead to delivery delays, resulting in uncertainty in the demand/supply of the current generation of rack-level AI servers. In this context, Goldman Sachs has lowered the target prices of Quanta, Hon Hai, FII, Wistron, AVC, and Auras by 7% to 21%.
<blockquote>
<img src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/1eb13b19-60ab-4a96-a668-839d3f090d9e.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg" alt="" original-src="https://imageproxy.pbkrs.com/https://wpimg-wscn.awtmt.com/1eb13b19-60ab-4a96-a668-839d3f090d9e.png"/>
</blockquote>
For investors, this market adjustment reflects a shift in the AI server industry from frenzied growth to rational expansion, and the valuations of supply chain companies have been adjusted accordingly. Goldman Sachs' analysis states that although growth is slowing, investment in AI infrastructure remains a major growth driver for the technology sector, but due to various limiting factors, its growth pace will be more moderate than previously expected by the market.
Risk Warning and Disclaimer
The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this is at your own risk

Goldman Sachs lowers global AI server shipment forecasts, reducing sales for 2025 and 2026 from 31,000 units and 66,000 units to 19,000 units and 57,000 units, respectively. This adjustment is due to the product transition period and supply-demand uncertainty. Goldman Sachs also lowered the target prices for related supply chain companies, with declines ranging from 7% to 21%, and downgraded Quanta Computer's rating from "Buy" to "Neutral." Nevertheless, Goldman Sachs remains optimistic about the market growth for AI training servers

- Goldman Sachs has revised its forecast for rack-level AI server shipments, lowering expected deliveries for 2025 and 2026 to 19,000 and 57,000 units, respectively.  
- This adjustment is attributed to product transition impacts and uncertainties in supply and demand, leading to a more conservative outlook for overall shipments.  
- Consequently, Goldman Sachs has also reduced target prices for several Taiwanese ODM and cooling supply chain companies by 7% to 21%, reflecting a shift from rapid growth to more measured expansion in the AI server market.

Wallstreetcn

Goldman Sachs

Goldman Sachs significantly lowered the global AI server shipment forecast, reducing the price expectations for corresponding supply chain stocks across the board