U.S. Industrial Production: Tech Capacity Expands as Consumer Output Weakens

August 2026 industrial production reveals a sharp divide: consumer-facing output and automotive products are weak, while computer and electronic products continue to grow and the capacity to make them is expanding. That contrast matters beyond the next trading session. It points to where the U.S. industrial system is allocating productive resources, even as the wider economy faces inflation risk.

The Federal Reserve’s G.17 release, published September 18, reports that total industrial production was unchanged in August and 1.4% above August 2025. Manufacturing output fell 0.3% on the month. The figures below are seasonally adjusted year-over-year changes through August 2026, unless a different period is stated. Output measures production; capacity measures the estimated ability to produce. They answer different questions.

Consumer production lags equipment

Source: Federal Reserve G.17 Table 1 and Table 9. Table 1 supplies the year-over-year production rates; Table 9 presents gross values of final products in billions of inflation-adjusted 2017 dollars. These are related measures, not interchangeable totals.

The contrast does not show a broad, tariff-driven manufacturing boom. It also does not establish that tariffs caused weakness in autos or consumer goods: these data alone cannot isolate the effect of trade policy from demand, financing costs, supply chains or other forces. The narrower finding is that equipment output is advancing much faster than consumer output. For background on the wider policy question, see our analysis of industrial production and tariffs.

Where the manufacturing growth sits

Source: Federal Reserve G.17 Table 1 and monthly supplement, Table 1D. Detailed subsector rates are rounded calculations from the seasonally adjusted series; the August observations are preliminary and may be revised.

This is the strongest near-term evidence for the technology thesis. Within durable manufacturing, electronics and communications are growing substantially faster than the aggregate. The August monthly detail is less uniform: the Fed reported a 0.5% monthly decline in computer and electronic products, and semiconductor output edged down from July. The year-over-year comparison is strong, but a straight-line monthly acceleration would overstate the evidence.

Capacity makes this a structural question

Industrial capacity is the Fed’s estimate of sustainable production capability. Rising capacity in a specific industry is more informative for a medium-term outlook than a single month of output, although an estimate of capacity is not a count of new factories or proof of future sales.

Source: Federal Reserve monthly supplement, Table 3. Percentages are calculated from the August 2025 and August 2026 capacity indexes shown there and rounded to one decimal place. The Fed’s separate fourth-quarter-to-fourth-quarter projections use a different window and should not be read as August year-over-year changes.

The semiconductor and communications figures suggest an industrial buildout with a longer horizon than the current monthly release. They are consistent with expectations of demand for computation, networking and related infrastructure. They do not isolate AI as the sole driver: defense, conventional telecom investment, replacement cycles, exports and other uses also contribute.

Electricity: supporting evidence, not a single-cause proof

Electric power generation, transmission and distribution (NAICS 2211) recorded +7.0% year-over-year output in August, while estimated capacity rose about 3.6% from August 2025. These are distinct measures: the output increase describes electricity-sector production; the capacity increase describes productive capability. Sources: Fed Table 1 and supplemental Table 3.

Power capacity is essential to large-scale data centers and other computing facilities, so its expansion fits the wider infrastructure picture. The electricity series covers the entire U.S. economy, however. Weather, residential and industrial demand, and electrification can all move it. It cannot by itself quantify electricity consumed by AI. Our earlier industrial production analysis provides a prior checkpoint for this developing theme.

Market implications: a long-horizon split

The structural signal is selective growth. Electronics output and capacity are expanding alongside power-sector capacity, while consumer goods and auto output remain under pressure. That combination supports a medium- to long-term watch on semiconductors, communications equipment, power infrastructure and the businesses that supply them. It is not a buy signal for an entire sector or an intraday trading setup: valuations, margins, financing and the ability to turn capacity into profitable sales still matter.

Inflation adds the counterweight. Higher energy, materials or import costs can make the buildout more expensive and keep interest rates restrictive, even when the underlying technology investment remains strong. Investors should watch subsequent Fed production and capacity revisions, utility demand, company spending plans and inflation data together. The next test is whether capacity expansion is sustained and matched by demand and returns.

Sources and methodology

• Federal Reserve G.17, September 18, 2026 — headline industrial production and capacity utilization.
• G.17 Table 1 and Table 9 — market and industry production and gross values.
• Monthly supplement Table 1D and Table 3 — detailed electronics output and capacity indexes.

Data through August 2026, seasonally adjusted where applicable; August output data are preliminary. The annual capacity comparison uses the Fed’s published August capacity indexes. Figures rounded to one decimal place. The annual revision scheduled for November 2026 may change historical readings.