Hyperscalers' Spending to Jump 90% This Year on AI Data Center Expansion

Nine Cloud Firms' Spending Set to Roughly Double Strong AI Demand Drives Higher Investment This Year's AI Server Shipment Forecast Raised Returns on Investment Diverge Across Companies

International|
|
By Park Yoon-sun
||
A view of Meta's Stanton Springs data center in Newton County, Atlanta, U.S., photographed on Jan. 13 (local time). AP Yonhap News - Seoul Economic Daily International News from South Korea
A view of Meta's Stanton Springs data center in Newton County, Atlanta, U.S., photographed on Jan. 13 (local time). AP Yonhap News

Total capital spending by the world's nine largest cloud service providers (CSPs) is projected to rise about 90% this year from a year earlier. The increase reflects a sharp rise in data center investment aimed at meeting growing demand for artificial intelligence (AI).

On the 4th, Taiwan's United Daily News, citing data from TrendForce, reported that combined capital spending this year by the nine major CSPs — Google, Amazon, Meta, Microsoft (MS), Oracle, ByteDance, Tencent, Alibaba and Baidu — is expected to reach $886.7 billion (about 1,263 trillion won), up roughly 90% from a year earlier. About 90% of the nine companies' total capital spending was accounted for by the five largest North American hyperscalers. The United Daily News explained that this "shows that major CSP companies are continuously expanding AI data centers to respond to computing power demand driven by the rapid growth of generative AI and large models."

The nine CSPs' combined capital spending next year is projected to expand further to $1.3 trillion (about 1,849 trillion won). While the growth rate of about 50% is lower than this year's, the United Daily News added that this stems from base effects and does not signal a slowdown in the flow of AI investment.

For example, TrendForce raised its forecast for AI server shipment growth this year to 31% from the previous 28%. This is because hyperscale CSPs and smaller data center operators have recently shown significantly higher intent to purchase AI server solutions. The forecast was also raised because the next-generation application-specific integrated circuit (ASIC) platforms of Google and Amazon Web Services (AWS) will enter mass production sequentially starting in the second half of this year, and Chinese companies are expanding their adoption of in-house AI solutions to meet cloud large language model (LLM) demand. Taiwan's securities industry projected that Taiwanese original design manufacturing (ODM) companies including Foxconn, Quanta, Wistron, Wiwynn and Inventec would benefit from the shipment increase.

Same Hyperscalers, Diverging Fortunes

Reuters Yonhap News - Seoul Economic Daily International News from South Korea
Reuters Yonhap News

Amid this, earnings fortunes among hyperscalers are also diverging.

In fiscal fourth-quarter (April-June) results disclosed by MS late last month, revenue rose 18% from the same period a year earlier to $90 billion (about 130 trillion won), significantly exceeding the market estimate of $87.62 billion. Net profit also increased 31% to $35.8 billion, with earnings per share (EPS) of $4.74, beating expectations of $4.24.

Revenue from Azure, MS's cloud business, surpassed $100 billion for the first time. Its quarterly growth rate reached 43%, up from the previous quarter's 40%. In addition to the cloud business, its AI platform "Copilot" also showed growth, with paying users rising from 20 million to 30 million in three months.

Google and Meta, on the other hand, showed a different trend. Meta's quarterly net profit came in lower than expected, and its free cash flow plunged more than 90%, sending its share price sharply lower after the earnings announcement. Alphabet, Google's parent company, also said its quarterly free cash flow turned to a loss for the first time since its listing.

MS's free cash flow also fell 23% from a year earlier to $19.6 billion due to increased capital spending, but it exceeded the market estimate of $13.4 billion. MS said its order backlog as of the end of the quarter stood at $678 billion, up $51 billion from the previous quarter. This provides grounds supporting the need to expand data center construction to respond to computing demand from AI companies such as OpenAI and Anthropic. While all three hyperscalers are pouring more than $100 billion a year into AI investment, MS, unlike Google and Meta, is assessed as having justified the need for its AI investment.

Original reporting by Park Yoon-sun for Seoul Economic Daily.

AI-translated from Korean. Quotes from foreign sources are based on Korean-language reports and may not reflect exact original wording.

Watch · Seoul Economic Daily

More →
2:32

AI KEY

Preview
Korean Corporate Intelligence HubKOSPI · KOSDAQ · 12 sectors

A live, cap-weighted view of every KOSPI and KOSDAQ sector, with same-day Korean reporting distilled by company — built for foreign investors, correspondents and analysts who need to scan Korea before the next session.

Korea Company Atlas

Preview
Market Ontology · The Feedback LoopKFTC 2025 · 92 groups · 121,954 articles

An English ontology of the Korean market — how companies, the media, the government and the National Assembly move each other in a loop. Korea's named controlling persons and designated business groups are a mechanism, not a risk to be priced blind.

SIGNAL

Pre-register
English Edition · Capital MarketsM&A · IPO · PE · Fund Flows

Pre-register for SIGNAL English Edition — a premium subscription bringing Korean capital markets coverage (M&A, IPOs, private equity, fund flows) to global institutional investors. First access to the 50% introductory rate.