Morgan Stanley discusses AI capital expenditure: Cloud giants have "deep pockets" and can further invest by 2025

The AI ​​boom can continue next year, which has always been one of the market's focus. Morgan Stanley recently analyzed the cash flow of large American cloud service providers in a research report, predicting that such companies will still have the ability to maintain sufficient capital expenditures next year and continue to invest in data centers for artificial intelligence.

According to the report, Morgan Stanley predicts that by 2025, the operating cash flow of the top four super-large cloud computing companies in the United States will reach $176 billion.

The report also analyzes the percentage of depreciation in total expenses for large cloud service providers. The report states that in recent years, due to heavy investments in data centers, the proportion of depreciation in total expenses of data center clients has significantly increased, from 3% in 2012 to 7%, rising to 5% to 10% by 2023. Morgan Stanley expects that with increased capital expenditures by the top four super-large cloud computing companies in the United States, depreciation will continue to grow, reaching 6% to 14% by 2025.

In addition, the report shows that the average AI capital expenditure/EBITDA for major enterprises in 2024 is around 40%. Morgan Stanley's analysis indicates that this suggests that the financial condition of super-large cloud computing companies is still healthy enough to have sufficient funds for further expenditures.

Morgan Stanley's cloud capital expenditure tracking index now shows that the year-on-year growth of this index in 2024 will reach 52% (excluding Amazon), higher than the previously predicted 49% year-on-year growth. It is expected that cloud capital expenditures between 2024 and 2025 will be close to $380 billion, nearly $50 billion more than the total of the past three years.

Morgan Stanley has updated the relevant data on AI server procurement in 2024. Overall, Morgan Stanley expects that super-large cloud computing companies will be more proactive in AI server procurement. However, Morgan Stanley also observed that Microsoft has reduced its allocation to AI in its capital expenditures. This is consistent with Aspeed's observation, as Aspeed previously pointed out that in the first half of 2024, some American cloud service providers' customers purchased more baseboard management controllers (BMCs) for traditional servers.

All major manufacturers are ready to increase investment

The report also summarizes the capital expenditure attitudes of major cloud service providers, with most companies stating that they have the ability to increase investment efforts For example, Google expects its quarterly capital expenditures for the full year to remain roughly at or above the $12 billion level in the first quarter. The company believes that the AI field is currently in a very transformative early stage in terms of technology, where the risk of underinvestment far outweighs the risk of overinvestment.

"We will continue to invest in designing and building robust and efficient infrastructure to support our efforts in AI, as we see many opportunities in the future."

Microsoft, on the other hand, stated that expenses related to cloud computing and AI account for almost all of the company's total capital expenditures, with capital expenditures for the 2025 fiscal year expected to be higher than in the 2024 fiscal year.

Meta believes that having sufficient computing power is crucial for many of these opportunities, so the company is heavily investing in infrastructure, raising its full-year 2024 capital expenditure forecast to $37 billion to $40 billion, up from $35 billion to $40 billion. It is currently expected that capital expenditures in 2025 will increase significantly as the company invests to support AI research and product development efforts.

Amazon also expects higher capital investments in the second half of 2024, with most of the spending going towards supporting the growing demand for AWS infrastructure, as the company sees strong demand in the AI field