How AI is Affecting GDP Growth, Productivity, and Jobs

Current Impact on U.S. GDP Growth

• AI investment contributes approximately 0.1 percentage points to the current U.S. GDP growth rate of roughly 2%.
• Without AI investment, the measured growth rate would be approximately 1.9%.
• Goldman Sachs estimates the contribution is far lower than the 0.5 to 1.0 percentage point figures cited by some economists.
• Measurement Factors:
  • A significant portion of AI hardware (e.g., semiconductors, servers) is imported; these costs are deducted as imports from U.S. GDP calculations.
  • Semiconductors are classified as intermediate inputs rather than final goods, further excluding them from standard GDP measures.
• Primary Drivers: Current U.S. GDP growth is driven more by consumer spending and non-AI equipment/IP investment than by AI.

Capital Expenditure (CapEx) Estimates

• Current Spending: Estimated at $600 billion in the U.S. and $1 trillion globally.
• Future Projections (2024–2030): Total cumulative investment is projected to range between $5 trillion and $7 trillion.
• Historical Comparison:
  • Peak AI investment is expected to reach 2.5% to 3% of GDP by 2027–2028.
  • This range is consistent with historical technological booms (e.g., the 19th-century railway boom peaked at ~5% of GDP, though the economy was much smaller then).
  • Nominal dollar amounts appear unprecedented only due to overall economic growth, not relative to the economy's size.
• Measurement Challenges:
  • Estimates vary due to three imperfect methodologies: adjusting hyperscaler spending, analyzing ecosystem revenues (risking double-counting), or utilizing national income accounts.

Productivity Growth Outlook

• Projected Impact: AI is expected to increase the level of productivity by approximately 15% relative to a non-AI scenario.
• Timeline: This impact is projected to diffuse and manifest over a decade.
• GDP Growth Revisions:
  • Long-term U.S. GDP growth forecasts have been revised up from 1.75% (pre-2020/2022) to 2.3%.
  • Goldman Sachs anticipates this figure could rise further into the mid-2% range or slightly above.
• Offsetting Factors: Net productivity gains must account for reduced investment in other areas (due to CapEx reallocation), slower labor force growth, low birth rates, and reduced immigration.
• Perception Gap:
  • C-suite executives report high expectations for productivity gains.
  • Frontline workers report minimal time saved, potentially due to incentives to understate efficiency or an expectation of increased workload.
  • Goldman Sachs views the "truth" as likely falling between these two extremes.

Labor Market and Employment Impact

• Job Loss Estimates:
  • AI exposure correlates with the loss of approximately 10,000 to 15,000 jobs per month in specific sectors.
  • Job creation in other areas (e.g., data center construction) partially offsets these losses.
• Task vs. Job Displacement:
  • Approximately 25% of work tasks/hours are exposed to AI.
  • Most exposure represents time redeployment rather than job elimination.
  • Only 6% to 7% of exposed tasks are expected to result in direct job elimination.
• Unemployment Projections:
  • Goldman Sachs expects unemployment pressure to rise by 0.5 to 1.0 percentage points over a 10-year transition period.
  • The forecast unemployment rate 10 years from now remains near the current 4.1%, assuming new job creation compensates for losses.
• Sectoral Variation:
  • Adoption rates range widely, with high penetration in tech sectors and low penetration in manual industries.
  • Recent survey methodology changes (from "production use" to "any business function") artificially inflated adoption statistics from 10% to over 20%.
• Historical Context:
  • 60% of current U.S. jobs existed in occupations not present in 1940.
  • Disruption is expected to be gradual ("slowly over time") rather than a sudden shock, remaining below the unemployment spikes seen in average (2–3%) or deep (5–6%) recessions.