share_log

“重估阿里”价值的焦点:AI时代的阿里云

The focus of "Reassessing Alibaba's" value: Alibaba Cloud in the AI era.

wallstreetcn ·  10:08

Goldman Sachs expects that about 80% of the 380 billion capital expenditure will be used for AI Servers. With the increase in relevant capital expenditure, Alibaba is expected to achieve revenue growth of over 20% in the fiscal years 2026 and 2027. According to a communication meeting held by UBS Group with Alibaba's management, in the short term, Alibaba Cloud's profit focus will be on the B2B sector, providing MaaS solutions through its QWen model and leveraging the cross-selling advantages of existing Cloud Computing products. In the long term, the application of AI agents is expected to further enhance profit margins.

In the era of AI, Alibaba Cloud is not only seen as Alibaba's new growth engine, but also becomes a key factor for investors to reassess value. $BABA-W (09988.HK)$

UBS Group stated in its latest Research Report that as the Taotian Group's Business stabilizes and new businesses gradually improve, investors should begin to recognize Alibaba's potential in Cloud Computing and generative AI. The report stated:

(Alibaba) management believes that the strong growth of the current AI Cloud Business is primarily limited by capacity. With the reduction in inference costs and the decrease in demand for chips, management believes that AI technology has become more affordable, and with the acceleration of digital transformation among Chinese enterprises and small and medium-sized enterprises, the AI Cloud Business has encountered significant growth opportunities.

Goldman Sachs' latest Research Report also pointed out that against the backdrop of Alibaba's three-year capital expenditure target being raised and stock performance outperforming Large Cap, the focus of investors is primarily on the potential returns of anticipated capital expenditures.

Goldman Sachs analyst Ronald Keung's team expects Alibaba's increased investment in Cloud and AI infrastructure and the enhancement of domestic inference chip supply to be sufficient to support its revenue growth of over 20% in fiscal years 2026 and 2027. Among them, the AI-related revenue of Alibaba Cloud is projected to reach 29 billion yuan and 53 billion yuan in fiscal years 2026 and 2027 respectively, accounting for 20% and 29% of total revenue.

How will the 380 billion capital expenditure be spent?

In mid-last month, Alibaba announced that over the next three years, it will invest more than 380 billion yuan to build Cloud and AI Hardware infrastructure, exceeding the total of the past ten years (approximately 300 billion yuan). This also sets a record for the largest investment ever made by a private enterprise in China in the field of Cloud and AI Hardware infrastructure construction.

Since then, foreign Institutions have successively raised Alibaba's Target Price. Goldman Sachs believes that Alibaba Cloud in the AI era will become a key driver of Alibaba's growth, especially in the context of the rapid development of China's AI market.

Goldman Sachs expects that out of Alibaba's 380 billion yuan capital expenditure, about 80% will be used for AI Servers, with the remaining portion allocated to General Servers, Datacenter Infrastructure, and Other Business Sectors. This investment plan will significantly enhance Alibaba Cloud's AI computing capabilities, especially in the context of rapidly growing demand for AI training and inference.

Goldman Sachs points out that although China faces some uncertainties in high-end chip supply, the increase in Alibaba's capital expenditure since 2024 and the gradual improvement in domestic chip supply will be sufficient to support Alibaba Cloud in achieving over 20% revenue growth in the fiscal years 2026-2027.

Research shows that Alibaba's capital expenditure structure has undergone a significant transformation, with about 90% allocated to Cloud Computing equipment, while this ratio was only 40-50% previously. Among AI Server investments, approximately 73% is designated for GPU, and 27% for Other Hardware. For AI Server investment, taking eight H20 GPU Servers as an example, the profit margin for a single server can reach 20-25% EBIT (Earnings Before Interest and Taxes) and 65-70% EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization).

Goldman Sachs expects that Alibaba Cloud's AI-related revenue will reach 9.9 billion yuan, 29 billion yuan, and 53 billion yuan respectively in the fiscal years 2025-2027, with the proportion of total revenue increasing from 8% to 29%. This growth is primarily attributed to the efficient use of AI Servers and the proliferation of AI applications and agents.

Alibaba Cloud's pricing advantages in AI training and inference, especially under tight supply of high-end GPUs, will lead to higher profit margins.

Goldman Sachs expects that Alibaba Cloud's EBITDA profit margin will gradually increase from 34% in the fiscal year 2025 to 40% in the fiscal year 2027, while the EBITA profit margin will rise from 9.7% in fiscal year 2025 to 12% in fiscal year 2027. Although the increase in depreciation expenses will put some pressure on EBITA profit margins, the high profit margins of the AI Cloud Business will partially offset this impact.

Goldman Sachs believes that Alibaba Cloud's AI Cloud Business EBITDA profit margin will be between 65%-70%, significantly higher than the profit margins of traditional public cloud businesses, which will bring substantial profitability improvements to Alibaba Cloud.

In the AI application layer that investors are focused on, Alibaba has Qwen and Quark.

As the investment in AI infrastructure gradually completes, investors' focus is shifting from AI infrastructure to AI applications and agents. Alibaba has also made significant progress in this field:

Leading open-source models Qwen2.5-Max, Wan2.1 (AI video model) and the latest QwQ-32B (inference model based on Qwen2.5 32B) perform comparably to leading models like DeepSeek-R1-671B.

Released a new version of the Quark application, supported by Alibaba's inference model based on Qwen.

Launched its first server-grade CPU C930, which will begin delivery to customers in March.

Established a strategic partnership with Manus, focusing on implementing Manus's functionality on domestic models and computing platforms.

Goldman Sachs emphasizes that the performance of Alibaba Cloud's Qwen model in inference and generative AI will further solidify its leading position in the AI market.

Goldman Sachs maintains a "Buy" rating for Alibaba, with a 12-month target price of $160 (or 156 HKD for the Hong Kong stock), which is about 13.5% higher than last Friday's closing price.

UBS Group reiterated its 'Buy' rating and raised Alibaba's Target Price from $160 / HK$154 to $176 / HK$172 (TTG's Target PE has been raised from 7 times to 8 times).

In addition, with the rise of cost-effective and well-functioning open-source models/environments, the model layer is expected to become more commoditized, and the reduction in computing costs will drive higher AI adoption rates, particularly in enterprise application scenarios, which will support years of growth in Cloud Computing / Datacenter demand.

As Alibaba Cloud is the largest public Cloud Computing Service provider in China, its market share is about twice that of Huawei Cloud and Tencent Cloud. Goldman Sachs expects that with the further development of AI models, the public Cloud Computing market in China will continue to maintain healthy growth. Alibaba Cloud's leading position in AI computing, especially its technological advantages in generative AI and large language models, will give it a favorable position in future market competition.

Editor/rice

The translation is provided by third-party software.


The above content is for informational or educational purposes only and does not constitute any investment advice related to Futu. Although we strive to ensure the truthfulness, accuracy, and originality of all such content, we cannot guarantee it.
23
Comment Comment 6 · Views 164.2k

Comment(6)

Recommended

Write a comment
6

Statement

This page is machine-translated. Futubull tries to improve but does not guarantee the accuracy and reliability of the translation, and will not be liable for any loss or damage caused by any inaccuracy or omission of the translation.