[Into the World of AI] Even Big Tech Bets on Debt...$700 Billion Wagered on AI Infrastructure

Securing Data Centers and Semiconductors to Meet AI Demand
Focus on the Potential for Recouping Increased Investment Costs

Major U.S. big tech firms—including Google, Microsoft, Amazon, and Meta—are investing over $700 billion (approximately 1,026 trillion won) this year to respond to explosive demand for artificial intelligence (AI), focusing their spending on securing AI data centers and semiconductors. The market is now keenly observing how quickly these tech giants will be able to recoup their increased investment costs through service revenue.

Sundar Pichai, CEO of Google. Photo by Yonhap News.

According to Bloomberg on the 24th (local time), the combined capital expenditure of the four big tech companies—Google, Microsoft, Amazon, and Meta—is projected to exceed $725 billion this year. This represents a 77% increase from last year’s capital expenditure of $410 billion. It is believed that these companies are channeling their funds into AI data centers, graphics processing units (GPUs), proprietary AI chips, servers, and building up power infrastructure.

Alphabet, Google’s parent company, announced during its second-quarter earnings report on the 22nd that it had raised its capital expenditure outlook for this year from the previous range of $180–190 billion to $195–205 billion. Anat Ashkenazi, Chief Financial Officer, stated, “The scale of AI infrastructure investment next year will be significantly larger.”

Alphabet has decided to further expand investment in data centers, servers, and proprietary AI accelerators in response to increasing demand for the generative AI model Gemini and the growing need for Google Cloud services. As it is becoming difficult to accommodate surging AI demand with just its own data centers, Alphabet also plans to rent additional capacity from external data centers. In order to maintain a full-stack strategy covering everything from AI infrastructure to AI models and services, Alphabet aims to secure enough computing resources.

The scale of AI infrastructure investment by Amazon, Microsoft, and Meta—who are due to announce their respective earnings next week—is also a major concern for investors. Amazon stated during its most recent earnings call that it will maintain its plan for about $200 billion in capital expenditures this year and affirmed its commitment to continued investment in AI infrastructure. Andy Jassy, Amazon’s CEO, said, “We plan to invest about $200 billion this year to respond to new opportunities such as AI and semiconductors,” adding, “We expect a high return on investment over the long term.”

In its April earnings disclosure, Microsoft announced its intention to continue expanding Azure data centers and AI computing infrastructure—including GPUs and CPUs—based on capital expenditures of about $190 billion. During the same month, Meta also raised its capital expenditure forecast for the year to $125–145 billion and said it would further invest in building data centers and securing AI computing capacity.

Cooling system on the rooftop of a data center located in Oakland, California, USA. Photo by Yonhap News

OpenAI is pursuing “Project Camellia,” which involves building a 3.2-gigawatt (GW) data center in Effingham County, Georgia. OpenAI has raised its expected cloud and computing spending by 2030 from $600 billion to $750 billion. SpaceXAI is also searching for a large-scale data center site in Texas and plans to establish a facility comparable to the 1 GW-class AI computing infrastructure currently operating in Memphis.

According to a recent report from the independent research institution SemiAnalysis, outstanding AI infrastructure-related debt raised by U.S. cloud service providers, data center developers, and hyperscalers for building AI servers and data centers will exceed $7 trillion by 2029, and cumulative AI capital expenditures from 2024 to 2029 are expected to reach approximately $11.1 trillion.

Market watchers believe that the AI investment competition has entered a new phase. The key question now is how quickly companies can offset surging capital expenditures with revenue from cloud and generative AI services. In its report, Morgan Stanley noted that the core issue in the AI infrastructure market is shifting from “how much is being invested” to “how much debt the AI ecosystem can sustain.”

Charu Chanana, Chief Investment Strategist at Saxo Markets, said, “Investors will increasingly focus on how much cash needs to be reinvested to maintain competitiveness and whether AI revenues can outpace capital spending, amortization, and operating costs.”

This content was produced with the assistance of AI translation services.