US Markets
Home›US Markets›Sectors›AI compute pricing slips as power constraints reshape…
AI compute pricing slips as power constraints reshape infrastructure bet
Stansberry Research analyst Gabe Marshank estimates data center buildout could reach about $4.5 trillion by 2030, requiring roughly $1.7 trillion in annual compute revenue to justify the spending.
MarketBeat Ratings, citing analysis from Stansberry Research, said AI infrastructure investors may not have fully repriced for signs that compute pricing is already slipping.
The outlet frames the core issue as a potential mismatch between where capital is flowing, into data center construction, and where it may ultimately need to flow, into reliable power sources.
According to the piece, Stansberry Research analyst Gabe Marshank estimates a data center buildout of roughly $4.5 trillion by 2030 and argues the compute layer would need around $1.7 trillion in annual revenue to earn an adequate return on that capital.
For context, the analysis compares that compute-layer revenue requirement to an estimated roughly $1.0 trillion in revenue tied to the existing platform economy, using companies such as Alphabet, Microsoft, Amazon, Meta Platforms, and Oracle.