AI's contribution to US productivity is becoming visible, currently showing up more as increased output than layoffs.
AI summary card
AI's contribution to US productivity is becoming visible, currently showing up more as increased output than layoffs.
Morgan Stanley believes that since 2025, high AI-exposure industries have led the acceleration in US labor productivity, contributing 1.7 percentage points of the 2.4 percentage-point productivity growth from 4Q to 4Q 2025, driven mainly by output growth rather than labor substitution.
- Output-per-employee growth in high AI-exposure industries was clearly faster in 2025 than in medium- and low-AI-exposure industries, and the acceleration in productivity was more pronounced.
- The productivity improvement mainly comes from output growth; employment growth trends are similar across industries with different levels of AI exposure, indicating that labor substitution remains limited for now.
- The report constructs an industry-level output-per-employee indicator using BEA industry output data and BLS employment data, and classifies AI exposure using the method of Felten et al. (2021).
- After the Middle East conflict, financial conditions at one point were equivalent to about a 30bp increase in the federal funds rate; after the April 7 ceasefire announcement, financial conditions eased by 42bp.
- The US 1Q GDP tracking estimate was raised by 0.2 percentage points to 2.4%, but delays in construction spending and housing starts data make the estimate more uncertain than usual.
Report interpretation
Overview
This issue of the US Economic Weekly focuses on the relationship between AI and US productivity, while also covering financial conditions, tariffs, 1Q GDP tracking, inflation, housing, trade, durable goods orders, the FOMC, and the upcoming economic data calendar. The core conclusion is that the AI wave has indeed coincided with faster labor productivity growth in high AI-exposure industries, but the current evidence supports "faster output growth" more than "improving efficiency by cutting jobs."
Core views
The report argues that since the beginning of 2025, high AI-exposure industries have been leading productivity growth. Morgan Stanley estimates that in the four quarters through 4Q 2025, overall US productivity grew by 2.4 percentage points, of which high AI-exposure industries contributed 1.7 percentage points, up from a 0.7 percentage-point contribution in 2024. At the industry level, AI exposure is positively correlated with output-per-employee growth in 4Q 2025, with the regression shown in the chart as y = 3.1x + 2.2 and a t-statistic of 4.0 for AI exposure. The report emphasizes that accelerating output rather than labor substitution is the main driver, so the disruption to the labor market remains limited for now; however, this pattern could still change as AI adoption increases.
Analysis framework
The report first uses industry output and employment data to measure labor productivity, then divides industries into three groups by AI exposure-high AI exposure, medium AI exposure, and low AI exposure-to compare changes in 2025 output, employment, and output per employee. At the same time, the report uses a financial conditions index within the FRB/US framework to assess the equivalent policy rate impact of asset-price changes on economic activity, and mechanically aggregates monthly activity data to track US 1Q GDP.
Methodology notes
Industry output per employee
Uses BEA industry output data and BLS nonfarm employment data to construct an industry-level proxy for labor productivity, in order to observe changes in output, employment, and productivity across industries with different AI exposure.
AI exposure grouping
Based on the AI exposure standard in Felten et al. (2021), industries are divided into the top 25% AI exposure, middle 50% AI exposure, and bottom 25% AI exposure to compare productivity performance.
Equivalent change in the federal funds rate from financial conditions
The model incorporates the 10-year US Treasury yield, S&P 500 returns, corporate BBB credit spreads, dollar valuation, and oil prices, and aggregates them using the growth elasticities relative to the federal funds rate in the FRB/US model, interpreting the result as the basis-point change in the federal funds rate needed to produce a similar effect on economic activity.
GDP tracking estimate
The GDP tracking estimate differs from the officially published GDP forecast; it reflects a mechanical aggregation of monthly activity data that directly enter the BEA calculation, and the corresponding components are updated whenever new data affect the tracking subcomponents.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- US macroeconomyCore subject of research
- Strengths
- 1Q GDP tracking was revised up to 2.4%, with relatively strong consumption and nonresidential fixed investment, while AI-related productivity improvement supports potential growth.
- Weaknesses
- Delays in construction spending and housing starts data, along with tariffs, oil prices, and geopolitical shocks, increase estimation uncertainty.
- Comparison
- Morgan Stanley's GDP tracking estimate is above the Atlanta Fed's 1.2% tracking estimate and also close to the NY Fed's 2.3% nowcast.
- Risks
- Tariff policy uncertainty, higher oil prices, supply constraints, sticky inflation, and government data delays.
- High AI-exposure industriesMain contributors to productivity improvement
- Strengths
- Growth in output per employee and the acceleration in productivity both lead in 2025, with high-exposure industries contributing most of overall productivity growth.
- Weaknesses
- Employment growth has not accelerated significantly; if AI adoption deepens in the future, the risk of labor substitution could rise.
- Comparison
- Compared with medium- and low-AI-exposure industries, high AI-exposure industries show more pronounced output growth and productivity gains.
- Risks
- AI investment returns falling short of expectations, slower diffusion of adoption, and regulatory or labor-market frictions.
- US interest ratesVariable for financial conditions and policy path
- Strengths
- Financial conditions eased by 42bp after the ceasefire announcement, offsetting some of the earlier tightening caused by the conflict.
- Weaknesses
- The net tightening since the conflict began has mainly been driven by higher 10-year Treasury yields and dollar appreciation.
- Comparison
- The report interprets changes in financial conditions as equivalent changes in the federal funds rate, making them comparable with the effects of monetary policy.
- Risks
- Oil prices and inflation pressure may delay rate cuts, and the FOMC remains patient in a highly uncertain environment.
- US dollarComponent of the financial conditions index
- Strengths
- Dollar movements can affect future economic activity through the financial conditions channel.
- Weaknesses
- Dollar appreciation has been one of the main drivers of the net tightening in financial conditions since February 28.
- Comparison
- Relative to oil prices, the report argues that higher 10-year Treasury yields and dollar appreciation have been the main tightening factors.
- Risks
- Geopolitics, rate expectations, and tariff policy may increase dollar volatility.
- Oil pricesVariable for financial conditions and inflation risk
- Strengths
- Changes in oil prices are included in the financial conditions model and can be used to quantify their impact on economic activity.
- Weaknesses
- Rising oil prices weigh on tighter financial conditions and real spending, and may also add inflation pressure.
- Comparison
- The report says the impact of oil prices is a secondary driver, smaller than the effects of long-end rates and the dollar.
- Risks
- Middle East uncertainty, supply constraints, and rising input costs.
- US equity marketComponent of the financial conditions index and carrier of growth expectations
- Strengths
- S&P 500 returns are one of the variables in the financial conditions model, reflecting the impact of asset prices on future economic activity.
- Weaknesses
- The report does not provide stock ratings or industry investment recommendations; the equity implication mainly comes through the macro growth and financial conditions channels.
- Comparison
- The equity market, together with rates, credit spreads, the dollar, and oil prices, jointly determines financial conditions.
- Risks
- A renewed tightening in financial conditions, downward earnings revisions, and policy or geopolitical shocks.
Key data
- Contribution of high AI-exposure industries to productivity growth1.7pp / 2.4ppIn the four quarters through 4Q 2025, high AI-exposure industries contributed 1.7 percentage points of the total 2.4 percentage-point productivity growth.
- Contribution of high AI-exposure industries in 20240.7ppThe report notes that the contribution accelerated in 2025, above the 0.7 percentage-point contribution in 2024.
- Regression relationship between AI exposure and productivityy = 3.1x + 2.2;t-stat = 4.0The chart shows a positive correlation between industry AI exposure and the four-quarter year-over-year change in industry output per employee in 4Q 2025.
- Tightening in financial conditions after the Middle East conflictabout 30bpSince the Middle East conflict began on February 28, financial conditions have tightened by an amount equivalent to about a 30bp increase in the federal funds rate.
- Change in financial conditions after the April 7 ceasefire announcementeased by 42bpThe report says financial conditions eased by 42bp after the ceasefire announcement.
- US effective tariff rate in February 2026about 8.5%The report estimates that the effective tariff rate on US imports was about 8.5% in February 2026.
- Baseline tariff estimateabout 11%Replacing the IEEPA tariff with a 15% Section 122 tariff and accounting for the import mix brings the baseline tariff close to 11%.
- Core tariff estimate13-14%Excluding fuel, gold, and AI-related imports, the core tariff estimate is closer to 13-14%.
- 1Q GDP tracking estimate2.4%Strong March retail sales and upward revisions to prior months raised the 1Q GDP tracking estimate by 0.2 percentage points to 2.4%.
- 1Q consumption estimate1.5%Consumption was revised up from 1.1% to 1.5%; excluding about 0.4 percentage points annualized of residual seasonal price effects, underlying consumption growth is about 2%.
- 1Q core PCE price4.1% q/q annualizedThe report expects 1Q core PCE prices to rise 4.1% quarter-over-quarter annualized.
- 1Q headline PCE price4.3% q/q annualizedThe report expects 1Q headline PCE prices to rise 4.3% quarter-over-quarter annualized.
- Private domestic final purchases2.2% q/q annualizedConsumption plus investment grew 2.2% quarter-over-quarter annualized, 0.2 percentage points below the average of the past year.
- Nonresidential fixed investment6.6%The report says nonresidential fixed investment grew at a strong 6.6% pace, partly dependent on February commercial construction spending data.
- Expected March goods trade deficit$74bnThe report expects the goods trade deficit to narrow by about $10 billion to $74 billion.
- April manufacturing ISM tracking52.9Solid demand and continued expansion in output show resilience, but supply and price uncertainty tilt near-term risks to the downside.
- Expected March light vehicle sales16.0mnThe report expects March light vehicle sales of 16.0 million, below the February pace and slightly below the prior year.
Impact & implications
If AI-driven productivity gains are achieved mainly through output expansion rather than large-scale layoffs, the macro implication is more positive: faster income growth could support either labor income or corporate income while easing market concerns about AI's short-term impact on employment. For policy, financial conditions, oil prices, tariffs, and inflation will still affect the Fed's degree of patience; for asset allocation, AI-related capital deepening, long-end US rates, the dollar, credit spreads, and oil prices remain important variables for judging the growth and policy path.
Risks
- As AI adoption accelerates, the current pattern of "higher output rather than job substitution" may change.
- Tariff policy faces uncertainty over whether Congress will extend Section 122 after its 150-day term.
- Conflict in the Middle East, rising oil prices, and supply constraints could push up inflation and weigh on real spending.
- Government shutdown-related delays in construction spending and housing starts data make GDP estimates more uncertain than usual.
- Although the FOMC may maintain an easing bias, high uncertainty means policymaking will be more patient.
- While manufacturing demand remains solid, supply and price uncertainty tilt near-term risks to the downside.
What to watch
- Whether the subsequent trends in output, employment, and output per employee in high AI-exposure industries continue to diverge.
- Whether productivity gains will still come without clear labor substitution after AI adoption diffuses to non-AI-related industries.
- The combined impact on the financial conditions index from 10-year Treasury yields, the dollar, BBB credit spreads, the S&P 500, and oil prices.
- Progress in Section 122, Section 232, and Section 301 investigations, and whether the effective tariff rate continues to rise from 8.5%.
- Official 1Q GDP data, along with the delayed releases of construction spending and housing starts data.
- Whether the FOMC statement and press conference continue to emphasize patience, and where market questions focus between inflation and growth risks.
- April manufacturing ISM, light vehicle sales, home prices, the goods trade deficit, and durable goods orders.