Daily Alpha · Substack
· Premarket Alpha · by Buzzberg Research
Noah Smith argues AI already delivers large consumer value while its data-center buildout raises labor demand and financing pressure.
Themes on this desk
Consumer value
Survey research estimates $172 billion of annual US consumer surplus from AI, exceeding the combined run-rate revenue of the largest private labs.
Labor and crowding out
The buildout could require 500,000 construction and trades jobs by 2030 while scarce labor and higher rates crowd out other activity.
Data-center construction crowds out other activity via scarce labor and higher rates
Smith concedes data-center construction 'certainly crowds out some other forms of economic activity, by sucking up scarce labor, and by raising interest rates (which makes it harder to finance other projects)'—acknowledging the AI boom's role in soaring interest rates.
Crowding-out via higher rates is a direct transmission channel from AI capex to financing costs across the broader economy, a headwind for rate-sensitive sectors and a risk to the 'free lunch' framing.
Watch Whether AI-driven capex continues to correlate with elevated long rates and squeezed non-AI business investment.
Source →Brynjolfsson et al. estimate $172B annual US consumer surplus from AI
Smith cites Brynjolfsson et al.'s April survey-based estimate of $172 billion in annual US consumer surplus from AI—more than the combined run-rate revenue of Anthropic and OpenAI.
A large, measurable consumer surplus supports the view that AI adoption creates real economic value beyond vendor revenue, underpinning durable demand for AI services.
Watch Follow-up consumer-surplus estimates and household adoption surveys; a declining surplus estimate would weaken the 'AI delivers broad value' case.
Source →Goldman estimates 500,000 new construction and trades jobs needed by 2030
Smith cites Goldman Sachs estimating that 500,000 new construction and trades jobs will have to be added by 2030 to sustain the accelerating data-center buildout.
A quantified multi-year labor-demand pipeline implies persistent wage pressure in skilled trades and supports construction/industrial suppliers tied to the buildout.
Watch Monthly construction employment and construction wage data versus the 500k-by-2030 trajectory; a shortfall would signal buildout deceleration.
Source →