Morgan Stanley's operating-lease-inclusive projection put 2027 and 2028 AI infrastructure spending at $1.58 trillion and $1.81 trillion, totaling approximately $3.4 trillion.
Concluded·Day 247·3 sources
Morgan Stanley projects ~$3.4T in AI infrastructure spending over 2027-2028
The gist
Multiple major financial institutions project AI infrastructure investment will roughly double in scale over the next two years, spanning GPUs, networking, power, and data center capacity. Wells Fargo forecasts the top three hyperscalers will hold roughly 75% of compute capacity through 2027.
The full picture
Morgan Stanley forecasts hyperscalers and neoclouds will spend $1.58 trillion in 2027 and $1.81 trillion in 2028 on AI infrastructure when operating leases are included, totaling approximately $3.4 trillion over those two years. A separate Morgan Stanley projection, which excludes operating leases, puts 2027 capex at approximately $1.2 trillion and 2028 at $1.4 trillion, revised upward by 9% and 10% respectively from prior forecasts. Spending is projected to surge roughly 62% from $978 billion in 2026 to $1.58 trillion in 2027. Individual 2028 estimates from the operating-lease-inclusive projection include Google at $434 billion, Amazon at $314 billion, Microsoft at $310 billion, and Meta at $303 billion. Wells Fargo separately forecasts hyperscaler compute capacity will more than double from 49 gigawatts at end of 2025 to over 125 gigawatts by 2028, with Alphabet projected to lead at 35 gigawatts, followed by Amazon at 29 gigawatts and Microsoft at 27 gigawatts.
How it developed
Morgan Stanley raised its 2027 and 2028 hyperscaler capex forecasts by 9% and 10%, to approximately $1.2 trillion and $1.4 trillion respectively.
Wells Fargo published forecast projecting hyperscaler capacity to exceed 125 gigawatts by 2028.
Sources
Milk Road AI Twitter
Related
The daily email
Want this in your inbox?
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free