Google CEO Sundar Pichai gestures during a meeting with France’s President Emmanuel Macron on the sidelines of the AI Impact Summit in New Delhi on Feb. 19, 2026.
Ludovic Marin | Afp | Getty Images
Wall Street analysts estimate total AI capital expenditures could now climb above $1 trillion in 2027, following even bigger spending plans unveiled by the hyperscalers during Wednesday’s tech earnings.
Both Evercore and Bank of America placed 2027 capex in excess 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 supply and pricing increases,” analysts for Jefferies said in a Thursday note to investors.
This year’s spending projections were up across the board, with Google parent 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, according to a tally by Bank of America.
Tech CEOs are projecting confidence about their artifical intelligence investments as evidence of monetization, such as ramping cloud revenue, flows through to the latest earnings reports, but the amped spending is still generating skepticism among investors.
Amazon CEO Andy Jassy said the company is “confident in the long term capex investments we’re making,” projecting $200 billion in buildout for the year.
First-quarter cloud revenue for Alphabet surged 63% on the year, prompting about a 10% jump in its stock. CFO Anat Ashkenazi said Wednesday that capex plans are increasing to meet “robust demand.”
Investors are hunting returns
The overall cost of the AI buildout has been causing heads to spin, but analysts say they’re seeing flowthrough from investments to revenue as valuations and market caps surge.
“Cap-ex keeps climbing, but [return on investment] ROI is evident via ~$2 trillion backlog and accelerating cloud growth,” Jefferies analysts said. “Margin leverage holds for the hyperscalers despite AI investments, highlighting structural [operating expense, or] opex discipline.”
Confidence on monetization is particularly high for Alphabet where backlog growth is undergirding computing inventories and expansions.
The “backlog supports [the] cap-ex super-cycle,” Brian Pitz wrote for BMO Capital Markets on Thursday. “Google’s backlog nearly doubled quarter-over-quarter with a 400% annual increase to $462B.”
“The majority of the backlog is for core Google Cloud Platform contracts … and Google expects to recognize just over 50% of it as revenue over the next 24 months,” he added.
While Google’s cloud revenue impressed analysts, Meta’s expansion plans troubled investors, who wanted to see more pay off for the investments it has been making. Shares were recently down about 8%.
“Meta likely remains in the penalty box pending clearer capex ROI,” analysts with Jefferies wrote in a Thursday note.
The scompany spent $72 billion on capex in 2025, and is expecting to double that in 2026 to between $125 billion and $145 billion. That’s up from a prior range of $115 billion to $135 billion.
“We are increasing our infrastructure capex forecast for this year,” Meta CEO Mark Zuckerberg said Wednesday. “Most of that is due to higher component costs, particularly memory pricing. But every sign that we’re seeing in our own work and across the industry gives us confidence in this investment.”
Meta’s free cash flow has been dwindling, dropping to just $1.2 billion in the first quarter from $26 billion in the same period last year.
Bank of America analysts said they expect to see sales and free cash flow improving in 2026 throughout the sector, helping to support the spending.
Who benefits?
The sustained capex growth is good news for chipmakers and gear providers. They are customers to the hyperscalers, and analysts were taking note on Thursday. First-quarter earnings for CPU-maker Intel were especially strong as the AI buildout is requiring more than just graphics processing units, or GPUs.
There is “strong and growing demand for various custom [application-specific integrated circuits] programs (TPU, Trainium, Maia and MTIA),” according to Evercore analysts, who touted “an acute focus on agentic-AI as a key use-case which we believe will prove to drive a CPU renaissance over the next several years.”
RBC Capital Markets maintains positive ratings on Nvidia, Micron Technology, Marvell, Astera Labs, Arm Holdings, and Lattice Semiconductor.
“Strong capex trends should also bode well for sector-perform rated AVGO, AMD, SNDK, and INTC,” RBC said. “AI demand is driving double-digit growth in wafer fab.”
