How much value requires new infrastructure, data systems, robotics, training, and organizational change?
The general-purpose-technology literature argues intangible investment — process redesign, data, retraining — historically dwarfs spending on the technology itself; whether AI follows that pattern or diffuses more cheaply is the open question.
View on the map → · Open in Browse →
The buildout got measured: computing infrastructure doubled to ~1.4% of US GDP, Goldman put the 2026–2031 capex path near $7.6 trillion, and the Fed built indicators concluding the economy is in an infrastructure phase rather than a displacement phase. Anthropic's session data adds the human side: domain expertise, not coding skill, is the binding complement.
in a buildout phase rather than the onset of broad-based displacement
Import-adjusted measure of AI investment's GDP contribution; indicators point to a buildout phase.
~$7.6T of AI infrastructure capex over 2026–2031, with the sensitivity analysis that moves it by hundreds of billions.
Data center, hardware, and networking spending reached ~1.4% of US GDP in Q1 2026 — double the 2015–2022 average.
People decide what to build, and the agent decides how to build it.
Session data: domain expertise, not coding background, drives success with agents — human capital as binding complement.
$581.7B global corporate AI investment in 2025 (up 130%); adoption faster than PCs or the internet.