Productivity growth and AI are linked in every 2026 earnings call, but the official data are more measured than the narrative. The Bureau of Labor Statistics' revised second quarter figures, released September 3, 2026, show nonfarm business labor productivity rising at a 1.4% annual rate in the quarter and 2.2% over the year, with output up 2.5% and hours worked up only 0.2% [1]. Since the business cycle began in the fourth quarter of 2019, productivity has grown at a 2.1% annual pace, faster than the decade before the pandemic [1]. Unit labor costs rose just 1.4% over the year, and the labor share of output fell to 52.8%, the lowest in a series that starts in 1947 [1]. This note reviews what the BLS data show about productivity growth and AI so far, what they cannot yet show, and what the pattern means for wages, inflation and markets.
What the BLS productivity data show for Q2 2026
The revised release confirmed the preliminary estimate of 1.4% quarterly growth published August 6, with output revised to 1.7% and hours to 0.3% [1][5]. Hourly compensation rose 2.6% at an annual rate in the quarter, so unit labor costs rose 1.2% [1]. Real hourly compensation, which adjusts for consumer prices, fell 3.3% in the quarter because of the spring surge in energy prices [1]. First quarter productivity was revised to 0.8% [1].
| Nonfarm business sector | Q2 2026 (annual rate) | Q2 2026 vs Q2 2025 |
|---|---|---|
| Labor productivity | +1.4% | +2.2% |
| Output | +1.7% | +2.5% |
| Hours worked | +0.3% | +0.2% |
| Hourly compensation | +2.6% | +3.7% |
| Unit labor costs | +1.2% | +1.4% |
| Real hourly compensation | down 3.3% | down 0.1% |
Source: BLS Productivity and Costs, Second Quarter 2026, Revised [1]. Manufacturing productivity rose 2.4% in the quarter, revised up from 1.9%, as output rose 5.4% [1].
Output up, hours flat: the signature of a productivity cycle
The year over year decomposition is the key fact. Output grew 2.5% while hours grew 0.2%, so almost all of the growth came from producing more per hour rather than from more labor [2]. That is consistent with the payroll data in DataPorium's economic metrics, which show total nonfarm employment of 159.1 million in August 2026, slightly below the 159.5 million of August 2025, while real GDP in the second quarter was 2.1% higher than a year earlier [4]. An economy that grows 2% with flat employment is, by definition, a productivity story.
Is AI showing up in the numbers yet?
The honest answer is that the BLS data cannot attribute productivity gains to any single technology. Labor productivity is output per hour; it rises when firms invest in equipment, software and organization, when the mix of jobs shifts toward higher value work, or when weak firms exit. What can be said is that the pattern since 2019, 2.1% annual growth versus roughly 1.5% in the prior decade, coincides with the largest business investment cycle in data centers and software in U.S. history, and the Federal Reserve's July 29 statement described productivity growth and capital investment as strong [1][3].
Three features of the data are consistent with an early AI effect. First, the gains are concentrated in output per hour rather than in hours, which is what a labor saving technology produces. Second, unit labor costs are rising slowly at 1.4% despite 3.7% compensation growth, meaning firms are absorbing wage increases through efficiency [1]. Third, the labor share has fallen to a record low 52.8%, which is what happens when returns accrue first to capital and to the owners of new technology before competition passes them to workers and consumers [1].
What the data cannot yet show
- Whether the 2019 to 2026 acceleration reflects AI or the earlier wave of cloud computing, remote work and pandemic driven business exits.
- How much of measured output growth is data center construction and chip investment itself, which counts as output today regardless of future returns.
- Whether productivity gains will broaden beyond software, finance and professional services into health care, education and construction, where measured productivity has long lagged.
Why productivity matters for wages, inflation and the deficit
Productivity is the only durable source of real wage growth. Hourly compensation rose 3.7% in the year to the second quarter; with productivity up 2.2%, unit labor costs rose 1.4%, which is below the Fed's 2% inflation goal and means labor costs are not a source of inflation in 2026 [1]. The inflation that remains, 3.4% on the CPI in July, is energy driven, and the CPI index level of 332.8 in July 2026 was 3.3% above a year earlier on a seasonally adjusted basis [4]. If productivity holds near 2% and energy prices settle, real wages can rise without any tightening of the labor market.
Productivity also matters for public finances. Federal debt sustainability depends on the gap between the average interest rate on the debt and nominal GDP growth. Every extra half point of trend productivity growth raises nominal GDP, tax receipts and the denominator of the debt ratio at the same time. From a market oriented perspective, the policy priority is to keep the investment cycle going: full expensing of capital spending, permitting reform for power and data center capacity, and a stable tax code do more for long run growth than any demand side stimulus.
The fair counterpoint is that a record low labor share and flat employment are uncomfortable in the short run, and that the productivity gains so far are modest compared with the late 1990s, when growth ran near 3% a year. Investors may consider that the current combination of 2% productivity growth, 1.4% unit labor cost growth and a falling labor share is unusually favorable for corporate margins, and that the main risk is a policy or energy shock that interrupts the investment cycle.
U.S. output rose 2.5% in the year to mid 2026 with hours up only 0.2%, so nearly all growth came from productivity, and unit labor costs of 1.4% show that wages are being paid for by efficiency rather than passed on as inflation.
Key takeaways
- Nonfarm business productivity rose 1.4% at an annual rate in Q2 2026 and 2.2% over the year, with output up 2.5% and hours up 0.2% [1][2].
- Since Q4 2019 productivity has grown 2.1% a year, faster than the pre pandemic decade; manufacturing productivity rose 2.4% in the quarter [1].
- Unit labor costs rose only 1.4% over the year despite 3.7% compensation growth, and the labor share fell to a record low 52.8% [1].
- The BLS data cannot attribute gains to AI, but the pattern of output growth without hours growth and strong capital investment is consistent with an early effect [1][3].
- Sustained 2% productivity growth would support rising real wages without inflation and improve federal debt dynamics [1][4].
Frequently asked questions
What was U.S. productivity growth in the second quarter of 2026?
Nonfarm business labor productivity rose at a 1.4% annual rate in Q2 2026 and 2.2% compared with Q2 2025, according to the BLS revised release of September 3, 2026 [1].
Is AI increasing productivity according to official data?
The BLS does not attribute productivity to specific technologies, but productivity has grown 2.1% a year since late 2019 with output rising 2.5% and hours only 0.2% over the past year, a pattern consistent with labor saving investment [1][2].
What are unit labor costs and why do they matter in 2026?
Unit labor costs are compensation per unit of output; they rose 1.4% in the year to Q2 2026 because 3.7% compensation growth was mostly offset by 2.2% productivity growth, which means labor costs are not driving inflation [1].
What does a record low labor share mean?
The labor share, the portion of output paid to workers as compensation, was 52.8% in Q2 2026, the lowest since 1947, meaning a larger share of the gains from higher output is currently going to capital rather than wages [1].
Sources & References
- [1] Bureau of Labor Statistics: Productivity and Costs, Second Quarter 2026, Revised
- [2] BLS The Economics Daily: Productivity up 2.2 percent from second quarter 2025 to second quarter 2026
- [3] Federal Reserve: FOMC Statement, July 29, 2026
- [4] DataPorium Economic Indicators and Macro Data
- [5] Bureau of Labor Statistics: Productivity and Costs, Second Quarter 2026, Preliminary
