Big Tech is spending money on AI systems like there’s no tomorrow. These massive investments are their weapon to win the AI race. Lately, the focus has been on one area in particular: data centers.
Data centers of power. These facilities house thousands of servers where information is received, stored, processed, and sent. To function, they need a lot of hardware and power. In other words, enormous amounts of money.
The more (and more distributed), the better. Microsoft is a prime example of this race to expand data centers. In the last 12 months, it has invested more than $30 billion and signed agreements with countries worldwide to create new data centers or boost AI systems development locally. Like its competitors, it aims to gain the upper hand in this AI battle, even as costs continue to rise. To understand the scale of this spending, look at capital expenditures (CapEx).
What’s CapEx? CapEx refers to a company’s spending on acquiring, improving, and maintaining long-term assets such as equipment, infrastructure, software, and real estate critical to its operations. Unlike operating expenses (payroll, utilities), CapEx reflects a company’s investments in future growth.
Combined CapEx approaches $60 billion. According to The New York Times, Big Tech companies have ramped up investments in generative AI technology. A review of the CapEx spending of five key players—Amazon, Apple, Google, Meta, and Microsoft—reveals their combined expenditures neared $60 billion in the second quarter of 2024, up from less than $50 billion in the prior quarter and $35 billion a year ago. This trend shows no signs of slowing.
While CapEx has been growing steadily, the explosion of AI systems has driven a sharp spike since mid-2023, fueled by the generative AI boom. Source: The New York Times.
More money, more challenges. All major tech companies are dedicating significant resources to AI systems. CapEx reports reveal these investments, primarily for AI-related data centers, although some spending supports other business areas, such as cloud infrastructure.
- Meta. At the start of 2024, Meta CEO Mark Zuckerberg announced plans to invest $30 billion in new infrastructure. In April, the figure reached $35 billion. By midyear, that estimate rose to $37 billion, with $14.5 billion already spent in the first six months. CFO Susan Li said that Meta expects “significant CapEx growth in 2025 as we invest to support our AI research and our product development efforts.” Meta’s ambition is also to create an artificial general intelligence (AGI).
- Google. This company has spent over $25 billion in CapEx in the first half of 2024. When asked if the company might be overspending, CEO Sundar Pichai dismissed the concern, saying, “The risk of underinvesting is dramatically greater than the risk of overinvesting.” Pichai argued the infrastructure could serve other uses even if AI systems development slows.
- Amazon. The e-commerce giant has shifted its focus to AI models, committing $100 billion over the next decade. In the first half of 2024, it spent $32.6 billion on CapEx, with expectations for even higher expenditures in the latter half. Amazon SVP and CFO Brian Olsavsky said: “The majority of the spend will be to support the growing need for AWS infrastructure as we continue to see strong demand in both generative AI and our non-generative AI workloads.”
- Microsoft. When it comes to AI-focused data centers, Microsoft stands out as a major player. While its combined CapEx for the first half of 2024 trails Amazon at $25 billion, the company has ambitious plans to ramp up spending significantly. These massive investments in data centers are so extensive they may hinder Microsoft’s ability to meet its 2030 goal of becoming carbon-negative. As the backbone for OpenAI and ChatGPT—the leading generative AI chatbot—Microsoft shoulders the responsibility of sustaining a resource-intensive platform, requiring substantial ongoing investment.
- Apple. In contrast, Apple’s AI systems investments are modest. It allocated just over $4 billion in CapEx for the first six months of 2024, focusing on developing AI chips for data centers rather than rapidly building infrastructure.
Google’s CapEx Growth (2004-present). Google invested in infrastructure during the smartphone and video booms (YouTube), but recent spending dwarfs past figures. Source: FinanceCharts.
A historic investment. AI-driven CapEx spending has reached unprecedented levels. During the Web 2.0 boom, investments in data centers never approached the current scale. Even Amazon and Microsoft, leaders in cloud infrastructure, reported maximum quarterly CapEx of $5 billion and $4 billion, respectively, before 2020.
Amazon’s CapEx has hit new heights in the last four years, with a marked acceleration in the past 12 months—and the trend shows no sign of slowing. Source: FinanceCharts.
Even Amazon and Microsoft, the architects of the world’s two dominant cloud infrastructures (AWS and Azure), once operated with more modest investments. Before 2020, Amazon’s annual CapEx peaked around $5 billion, while Microsoft’s hovered near $4 billion. Since then, those numbers have surged—particularly over the past year.
A long-term gamble. CapEx reflects medium- and long-term investments, making them a window into strategic priorities. In the case of AI systems, Big Tech leaders are transparent about their extended timelines for profitability. Microsoft CFO Amy Hood told Forbes that spending on generative AI assets “will be monetized over 15 years and beyond.” It’s a clear pitch to investors: Patience will be rewarded. Whether that patience holds remains to be seen.
Spiraling out of control? Despite the optimism from tech executives, some analysts are skeptical about the sustainability of these investments. Many believe the AI systems boom may be veering into bubble territory. A Goldman Sachs report predicts that spending on AI models will continue to climb, potentially reaching $1 trillion in the coming years. As Jim Covello, a lead analyst on the report, put it, “What trillion-dollar problem is AI going to solve?” It’s a valid question—and one that Big Tech CEOs have yet to fully answer.
Image | Daniel Reynaga
View 0 comments