Citigroup has increased its forecast for AI-related infrastructure spending by tech giants to exceed $2.8 trillion through 2029, up from the previously estimated $2.3 trillion. This adjustment is due to aggressive early investments by hyperscalers and a growing appetite from enterprises.
The AI boom, sparked by the launch of ChatGPT in late 2022, has continued to drive significant capital expenditures and data center expansion. This is despite a brief crisis of confidence caused by China’s cheaper DeepSeek model and ongoing market concerns regarding U.S. President Donald Trump’s tariff policies.
The Wall Street brokerage forecasts AI capital expenditure across hyperscalers to reach $490 billion by the end of 2026, an increase from the earlier estimate of $420 billion.
Data center operators, or hyperscalers, such as Microsoft, Amazon, and Alphabet, have already invested billions of dollars to alleviate capacity constraints that have hindered their ability to meet the increasing AI demand.
Citi analysts noted that hyperscalers are likely to reflect this incremental spending in their third-quarter earnings calls, with guidance expected to be “building ahead of visible enterprise demand.” Citi estimates global AI compute demand will require 55 gigawatts of new power capacity by 2030, equating to $2.8 trillion in incremental spending, with $1.4 trillion in the U.S. alone.
The brokerage stated that big tech firms are no longer solely relying on profits to fund AI infrastructure. The costs are extremely high, about $50 billion for every 1 GW of compute capacity, and companies are borrowing to keep up.
This shift is already evident in their financials, as spending begins to impact free cash flows. Investors are now questioning how tech companies will fund this level of investment, particularly as traditional models fall short.
“Enterprises have provided a clear external validation of value,” Citi remarked, highlighting production deployments at companies such as Eli Lilly, Hitachi, and Wolters Kluwer.
