How large is the capital expenditure for AI?

Core Viewpoints

Recently, Alibaba announced an increase in AI investment, and other manufacturers may follow suit. We are paying attention to its macro scale.

Consider: 1) Alibaba is expected to invest 130 billion in AI by 2025, with an incremental increase of about 53.3 billion; 2) The three major telecom operators are expected to invest 100 billion in AI, with an incremental increase of about 20 billion; 3) Other companies do not have clear data. Assuming that Tencent and ByteDance's AI investment and incremental investment are on par with Alibaba, the total investment from six companies in AI by 2025 is expected to be about 490 billion, with an incremental investment of about 179.9 billion, which may drive fixed asset investment by about 0.35 percentage points in 2025. However, since some equipment needs to be procured from abroad, this estimate may be overestimated.

Report Summary

(1) Alibaba: Incremental investment may exceed 53.3 billion by 2025

Alibaba's capital expenditure exceeded expectations, with an estimated incremental AI investment exceeding 50 billion by 2025. On February 20, Alibaba announced its 2024 Q4 (calendar year) performance, with capital expenditure of 31.7 billion for the quarter, a quarter-on-quarter increase of 82%. According to Xinhua News Agency, "Alibaba Group CEO Eddie Wu announced on the 24th that in the next three years, Alibaba will invest more than 380 billion to build cloud and AI hardware infrastructure, totaling more than the past decade combined," which implies an average annual capital expenditure of 130 billion. According to Alibaba's financial report, its total capital expenditure for 2024 was 76.7 billion, indicating an incremental increase of about 130-76.7=53.3 billion for 2025.

Where will the investment go? Three major areas of AI. According to the introduction by Alibaba CEO Eddie Wu during the conference call, Alibaba's capital expenditure is mainly used in three major areas: infrastructure construction for AI and cloud computing, AI foundational model platforms, and AI-native applications and the AI transformation and upgrading of existing businesses.

(2) Three Major Telecom Operators: Expected Incremental Investment of About 20 Billion

By 2025, the total capital expenditure of the three major telecom operators may decline, mainly due to the drag from traditional businesses. According to the Communication Industry Report, the total capital expenditure of the three major telecom operators is expected to be 320 billion in 2025, a decrease of 14 billion compared to 2024, mainly due to the drag from traditional businesses. For example, mobile network investment is expected to total 106 billion in 2025, a decline of 10.5 billion from last year.

AI computing power investment is expected to grow by 20 billion compared to last year. According to the Communication Industry Report, "As the market demand for intelligent cloud, computing networks, intelligent computing services, and industry digital applications continues to rise, telecom operators will continue to increase investment in computing power, such as building new intelligent computing clusters, expanding the coverage of 400G computing power networks, accelerating the training and iteration of AI large models, and speeding up the integration of AI with industry digital applications." According to data from the Communication Industry Report, the three major telecom operators' computing power expenditure is expected to be 100 billion in 2025, an increase of 20 billion compared to 2024 (3) Other Companies: May Follow Suit

1. Tencent has recently actively introduced DeepSeek, and AI investment may increase. In February 2025, Tencent announced that DeepSeek-R1 would be integrated into platforms such as WeChat, Tencent Yuanbao, and QQ Music. With the increasing demand for computing power, Tencent may increase its investment in AI, but specific details will need to wait for Tencent's performance meeting around March.

2. ByteDance, as it is not publicly listed, relevant financial data is difficult to obtain. Considering the increasing traffic of Doubao, increasing AI investment is necessary.

3. Baidu, AI investment is a "long-term strategic focus," but the scale may be relatively small. Referring to Baidu's 2024 Q4 performance meeting, Baidu's acting CFO He Junjie stated, "We will continue to invest in enhancing artificial intelligence capabilities as a long-term strategic focus..." However, its capital expenditure scale may be small. According to 2024 data, Baidu's total capital expenditure was 8 billion yuan, only 10% of Alibaba's.

Report Directory

Report Body

1. How large is the scale of AI capital expenditure?

On February 20, Alibaba announced its quarterly performance, with capital expenditure significantly exceeding expectations. With the leading role of major players, other companies may also follow suit, and this article focuses on its macro scale.

(1) Alibaba: Investment increment in 2025 may exceed 50 billion

Alibaba's capital expenditure exceeded expectations, and it is expected that the AI capital expenditure increment in 2025 will exceed 50 billion. On February 20, Alibaba announced its 2024 Q4 (calendar year) performance, with capital expenditure of 31.7 billion yuan, an 82% increase from the previous quarter's 17.5 billion yuan. According to Xinhua News Agency, "Alibaba Group CEO Eddie Wu announced on the 24th that in the next three years, Alibaba will invest more than 380 billion yuan to build cloud and AI hardware infrastructure, totaling more than the past decade," which means an average annual capital expenditure of 130 billion yuan. Referring to Alibaba's financial report data, Alibaba's total capital expenditure in 2024 was 76.7 billion yuan, which may imply that the capital expenditure increment in 2025 is approximately 130 billion - 76.7 billion = 53.3 billion yuan.

According to Alibaba CEO Eddie Wu's introduction in the conference call, Alibaba's capital expenditure is mainly used in three major areas:

Ø1) Infrastructure construction for AI and cloud computing. The AI era has a clear and huge demand for infrastructure, and Alibaba will actively invest in AI infrastructure construction, with expected investments in cloud and AI infrastructure in the next three years surpassing the total of the past decade Ø2) Invest in AI foundational model platforms and AI native applications. The AI foundational large model is of great significance for the transformation of industry productivity, which will greatly enhance the R&D investment in AI foundational models, ensure technological advancement and industry leadership, and promote the development of AI native applications.

Ø3) Invest in the AI transformation and upgrading of existing businesses. For e-commerce and other internet platform businesses, the upgrade of AI technology will bring significant opportunities for enhancing user value. Therefore, continuous enhancement of R&D investment in AI applications and computing power will be made, using AI technology to deeply transform and upgrade various businesses, seizing new development opportunities in the AI era.

(2) Three major telecom operators: AI investment increment of about 20 billion in 2025

In 2025, the total capital expenditure of the three major telecom operators may decline, mainly due to the drag from traditional businesses. According to the Communication Industry Report, the total capital expenditure of the three major telecom operators is expected to be 320 billion yuan in 2025, a decrease of 14 billion compared to 2024, mainly due to the drag from traditional businesses. For example, mobile network investment is expected to total 106 billion yuan in 2025, a decline of 10.5 billion compared to last year.

AI computing power investment is expected to increase by 20 billion compared to last year. According to the Communication Industry Report, "As the market demand for intelligent cloud, computing network, intelligent computing services, and industry digital applications further increases, telecom operators will continue to increase investment in computing power, such as building new WanKa intelligent computing clusters, expanding the coverage of the 400G computing power network, accelerating the training and iteration of AI large models, and speeding up the integration of AI with industry digital applications." According to data from the Communication Industry Report, the computing power expenditure of the three major telecom operators is expected to be 100 billion yuan in 2025, an increase of 20 billion compared to 2024.


(3) Other manufacturers: may follow suit

1. Baidu, AI investment is a "long-term strategic focus," but the scale has not been mentioned. Referring to Baidu's Q4 2024 earnings briefing, when asked "What will be the strategic focus of Baidu's business investment?" Baidu's acting CFO He Junjie responded, "We will continue to invest in enhancing our artificial intelligence capabilities as a long-term strategic focus. On this basis, we will further deepen the AI transformation across our entire product line, especially in the search business." However, its capital expenditure scale may be relatively small; according to 2024 data, Baidu's total capital expenditure is 8 billion yuan, only 10% of Alibaba's.

2. Tencent is actively introducing DeepSeek, and AI investment may increase. In February 2025, Tencent announced the gray testing of the DeepSeek-R1 model in WeChat Search and simultaneously integrated it into platforms such as Tencent Yuanbao and QQ Music. With the increasing demand for computing power, Tencent is expected to increase its capital expenditure in the AI field; however, specific details will need to wait for Tencent's explanation at the earnings meeting around March.

3. ByteDance, as ByteDance is not publicly listed, relevant financial data is difficult to obtain. Considering the increasing traffic of Doubao, increasing AI investment is necessary.

Author of this article: Zhang Yu, source: Yi Yu Zhong De, original title: "AI Capital Expenditure, How Big is the Scale?" Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment goals, 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. Investment based on this is at one's own risk