Google CEO Sundar Pichai gestures throughout a gathering with France’s President Emmanuel Macron on the sidelines of the AI Affect Summit in New Delhi on Feb. 19, 2026.
Ludovic Marin | Afp | Getty Pictures
Wall Road analysts estimate complete AI capital expenditures may now climb above $1 trillion in 2027, following even larger spending plans unveiled by the hyperscalers throughout Wednesday’s tech earnings.
Each Evercore and Financial institution of America positioned 2027 capex in extra of a $1 trillion after the earnings calls, with 2026 estimates rising to between $800 and $900 billion.
“Cap-ex continues to soar as demand outpaces provide and pricing will increase,” analysts for Jefferies mentioned in a Thursday notice to traders.
This yr’s spending projections had been up throughout the board, with Google mum or dad Alphabet up 4% to $185 billion, Amazon up 1% to $200 billion, Meta up 8% to $135 billion, and Microsoft up a whopping 24% to $190 billion, based on a tally by Financial institution of America.
Tech CEOs are projecting confidence about their manmade intelligence investments as proof of monetization, corresponding to ramping cloud income, flows by means of to the most recent earnings reviews, however the amped spending continues to be producing skepticism amongst traders.
Amazon CEO Andy Jassy mentioned the corporate is “assured in the long run capex investments we’re making,” projecting $200 billion in buildout for the yr.
First-quarter cloud income for Alphabet surged 63% on the yr, prompting a couple of 10% leap in its inventory. CFO Anat Ashkenazi mentioned Wednesday that capex plans are rising to satisfy “sturdy demand.”
Traders are searching returns
The general value of the AI buildout has been inflicting heads to spin, however analysts say they’re seeing flowthrough from investments to income as valuations and market caps surge.
“Cap-ex retains climbing, however [return on investment] ROI is obvious through ~$2 trillion backlog and accelerating cloud progress,” Jefferies analysts mentioned. “Margin leverage holds for the hyperscalers regardless of AI investments, highlighting structural [operating expense, or] opex self-discipline.”
Confidence on monetization is especially excessive for Alphabet the place backlog progress is undergirding computing inventories and expansions.
The “backlog helps [the] cap-ex super-cycle,” Brian Pitz wrote for BMO Capital Markets on Thursday. “Google’s backlog almost doubled quarter-over-quarter with a 400% annual improve to $462B.”
“Nearly all of the backlog is for core Google Cloud Platform contracts … and Google expects to acknowledge simply over 50% of it as income over the subsequent 24 months,” he added.
Whereas Google’s cloud income impressed analysts, Meta’s growth plans troubled traders, who wished to see extra repay for the investments it has been making. Shares had been just lately down about 8%.
“Meta possible stays within the penalty field pending clearer capex ROI,” analysts with Jefferies wrote in a Thursday notice.
The scompany spent $72 billion on capex in 2025, and is anticipating to double that in 2026 to between $125 billion and $145 billion. That is up from a previous vary of $115 billion to $135 billion.
“We’re rising our infrastructure capex forecast for this yr,” Meta CEO Mark Zuckerberg mentioned Wednesday. “Most of that is because of greater element prices, significantly reminiscence pricing. However each signal that we’re seeing in our personal work and throughout the business offers us confidence on this funding.”
Meta’s free money stream has been dwindling, dropping to only $1.2 billion within the first quarter from $26 billion in the identical interval final yr.
Financial institution of America analysts mentioned they anticipate to see gross sales and free money stream bettering in 2026 all through the sector, serving to to help the spending.
Who advantages?
The sustained capex progress is sweet information for chipmakers and equipment suppliers. They’re clients to the hyperscalers, and analysts had been taking notice on Thursday. First-quarter earnings for CPU-maker Intel had been particularly sturdy because the AI buildout is requiring extra than simply graphics processing models, or GPUs.
There’s “sturdy and rising demand for varied customized [application-specific integrated circuits] packages (TPU, Trainium, Maia and MTIA),” based on Evercore analysts, who touted “an acute deal with agentic-AI as a key use-case which we consider will show to drive a CPU renaissance over the subsequent a number of years.”
RBC Capital Markets maintains optimistic rankings on Nvidia, Micron Expertise, Marvell, Astera Labs, Arm Holdings, and Lattice Semiconductor.
“Sturdy capex traits also needs to bode effectively for sector-perform rated AVGO, AMD, SNDK, and INTC,” RBC mentioned. “AI demand is driving double-digit progress in wafer fab.”
