Wall Street Now Sees AI Infrastructure Spending Topping $1 Trillion by 2027

The numbers coming out of Big Tech earnings this week have analysts revising their pencils. Evercore and Bank of America both now estimate that total capital spending on AI infrastructure could cross the $1 trillion mark in 2027, with 2026 projections landing between $800 billion and $900 billion.

Jefferies analysts noted in a Thursday note that spending keeps climbing because demand is outstripping supply and prices are rising. This year alone, Alphabet raised its capex forecast 4% to $185 billion, Amazon went up 1% to $200 billion, Meta jumped 8% to $135 billion, and Microsoft surged 24% to $190 billion, according to Bank of America’s tally.

CEO confidence is high. Amazon’s Andy Jassy said the company is “confident in the long term capex investments we’re making,” while Google CFO Anat Ashkenazi pointed to “robust demand” as the reason for increased spending. Alphabet’s cloud revenue jumped 63% year-over-year in Q1, sending its stock up about 10%.

Jefferies sees clear return on investment, citing a roughly $2 trillion backlog and accelerating cloud growth. BMO’s Brian Pitz highlighted Google’s backlog nearly doubling quarter-over-quarter to $462 billion, with over half expected to convert to revenue in the next 24 months.

Not everyone is convinced. Meta’s shares dropped about 8% as investors wanted more proof of payoff. The company spent $72 billion on capex in 2025 and expects to double that in 2026. Free cash flow fell to just $1.2 billion in Q1, down from $26 billion a year ago.

For chipmakers and gear suppliers, the sustained spending is a tailwind. Evercore pointed to strong demand for custom chips like TPU, Trainium, and Maia, while RBC maintained positive ratings on Nvidia, Micron, Marvell, and others, noting that AI demand is driving double-digit growth in wafer fabrication.

Source: CNBC