Four Tech Giants' Capital Expenditures Exceed $1.1 Trillion Since 2023
Google, Amazon, Microsoft, and Meta have collectively spent $1.1 trillion on capital expenditures from the start of the AI boom in 2023 until the end of June 2026.
The four companies plan to invest approximately $745 billion this year, primarily for data centers, advanced chips, and power infrastructure; they are also committing to a large number of long-term leases and procurement agreements.
The event-driven AI infrastructure arms race is accelerating, with funds continuously flowing into computing power and energy construction, benefiting the chip, power, and construction supply chains, while putting pressure on the tech giants' free cash flow and shareholder return expectations.
Source: Public Information
ABAB AI Insight
Since the AI wave triggered by ChatGPT, the four major cloud providers have continuously raised their capital expenditure guidance, with cumulative spending surpassing a trillion dollars, and new contractual obligations nearing $900 billion in a single quarter.
The capital path focuses on data centers, GPUs, and power, motivated by the need to seize high ground in training and inference capacity, while locking in future supply through long-term leases to avoid being outpaced by competitors.
A similar wave of collective capital expenditure was seen around the year 2000 during telecom fiber construction and in the 2010s during cloud computing expansion; currently, AI infrastructure is in a phase of "must invest or fall behind."
This essentially represents capital concentration: a few tech giants monopolize AI computing power investment, with high fixed costs and economies of scale forcing continuous increases, creating a winner-takes-all infrastructure moat.
ABAB News · Law of Cognition
The ticket to the AI arms race is trillion-level investment.
Cash flow for capacity is the new normal.
Infrastructure monetization precedes application monetization.