Atlas 2034 — projected US labor gap by metro and occupation
Atlas 2034 · long-run projection
~9.83M
additional workers needed by 2034,
if wages rise no more than
10%
above the baseline trend over the decade.
AI scenario AI scenario. Three intensities of AI deployment, drawn from Autor & Thompson (NBER w33941, 2025).
- Baseline: no AI-specific shock — the reference point.
- Gradual: modest cognitive substitution + modest physical complementarity.
- Transformative: aggressive cognitive substitution (Computer/Math, Business, Legal demand drops sharply) + large physical-work demand increases (Production, Transportation). Produces the U-shape headline result.
Immigration Immigration. Three supply-side immigration paths in the cohort-flow model. Cross-occupation and cross-metro structure is held fixed; only inflow magnitude changes.
- Baseline: historical (2010–2024 ACS) immigration patterns continue.
- Lower: all immigrant inflows by metro × education × occupation are halved.
- No new: zero net immigration over the decade.
Wage-growth cap (10-yr) Wage-growth cap (10-yr). Cap on the EXTRA % wage increase above the baseline trend over the decade. Wages are growing regardless; the slider limits how much further they can rise to clear the shortage.
The "shortage" you see on the map is whatever demand the model couldn't clear at this cap — wages literally can't rise far enough to attract more workers into those occupations and metros. Lower the cap → bigger reported shortage, because the model attributes more of the unmet demand to the wage constraint.
The "shortage" you see on the map is whatever demand the model couldn't clear at this cap — wages literally can't rise far enough to attract more workers into those occupations and metros. Lower the cap → bigger reported shortage, because the model attributes more of the unmet demand to the wage constraint.
View
Scroll to zoom · drag to pan · click to drill in
Worker gap, 2034 Gap = 0 → Shortage
— —