Last checked
Short answer
US GDP grew 3.4% annualised in Q3 2026 (BEA advance estimate, 29 October), up from 2.2% in Q2. BEA credits investment, consumers and government, offset by lower exports. But inventories (1.61 points) and government (0.90, mostly because SPR oil sales, which BEA books as negative spending, were smaller than in Q2) added roughly 2.5 points, imports subtracted 2.64, and BEA's cleaner gauge of private demand grew 4.9%.
- What happened BEA's advance estimate on 29 October put Q3 2026 real GDP growth at 3.4%, above the 3.0% to 3.2% consensus range and the Atlanta Fed's final 3.3% nowcast.
- Why it matters Equipment investment jumped 37.4% for information processing gear, but much of that hardware was imported, and imports alone took 2.64 points off GDP.
- What is cleaner Final sales to private domestic purchasers grew 4.9%, versus 4.6% in Q2, which is a steadier picture than the headline jump from 2.2% to 3.4%.
- What to watch The second estimate lands 25 November; over the last five quarters the gap between advance and third estimates ranged from 0.1 to 0.9 points.
Key facts
| Q3 2026 real GDP (advance) | 3.4% annual rate, vs 2.2% in Q2 (BEA, 29 Oct 2026) |
| Final sales to private domestic purchasers | 4.9% in Q3, vs 4.6% in Q2 (BEA) |
| Information processing equipment | +37.4% (AP, citing BEA); imports +19.7%, exports −5.7% |
| Contribution of imports | −2.64 points; inventories +1.61; government +0.90 (StockTitan tabulation of BEA tables) |
| Quarterly PCE prices in Q3 | 2.1% (core 2.5%), down from 5.1% (core 3.4%) in the Q2 advance estimate |
| Next dates | Second estimate 25 Nov 2026, 8:30 a.m. EST; FOMC decision 9 Dec 2026 |
What did BEA report on 29 October?
The Bureau of Economic Analysis said real gross domestic product rose at a 3.4% annual rate in the third quarter of 2026. After its own revisions, the second quarter ended at 2.2%. The number beat the forecasts that outlets had collected beforehand: AP cited a FactSet consensus of 3%, Trading Economics had 3.2%, and the Atlanta Fed's GDPNow model closed on 28 October at 3.3%.
BEA attributed the increase to investment, consumer spending and government spending, partly offset by a drop in exports. The acceleration from Q2, in BEA's words, primarily reflected an acceleration in investment and an upturn in government spending.
According to AP's reading of the release, consumer spending grew 3.8%, business investment outside housing rose 12.7%, information processing equipment jumped 37.4%, software 9.6% and research and development 7.5%. Housing investment fell 3.3%, its sixth decline in the last seven quarters. Trading Economics put total private domestic investment at 19.4%, up from 4.6% in Q2, with exports down 5.7% and imports up 19.7%.
The White House called it the strongest pace in a year and the strongest quarter for business investment in nearly 15 years, with growth over the past three quarters at a 2.7% rate. Those comparisons are the administration's characterisation and we have not checked them against BEA's historical tables. BEA's own text describes the equipment gain as computers and peripheral equipment and does not attribute it to AI.
How much of the investment surge was imported?
The accounting matters here. BEA says the rise in equipment investment was led by information processing equipment, mainly computers and peripheral equipment. It also says the largest contributor to the rise in imports was non-automotive capital goods. Those categories overlap heavily, though BEA does not say by how much. A server bought by a US company from an overseas supplier counts as a plus in investment and a matching minus in imports. GDP measures domestic production, so the net effect of imported hardware on the headline is far smaller than the 37.4% equipment figure suggests.
EY-Parthenon sized the import side: imports of computers, peripherals and parts were up 62% from a year earlier, and 'other' imports, which include semiconductors, were up 32%. On the tabulation StockTitan drew from BEA's advance tables, business and housing investment together added 1.66 points to growth, while imports subtracted 2.64 and falling exports another 0.68, for a trade drag of 3.32 points. GDPNow's final estimate had net exports subtracting 2.95 points, so the direction was no surprise to the model, even if the size was.
The equipment surge is real as a measure of what US firms bought, but the trade line took back more than the investment line gave. With imports and exports combined into one trade line, the five contribution lines sum to 3.37, consistent with the 3.4% headline. Fixed investment added 1.66 points and trade subtracted 3.32, so consumer spending (+2.52), inventories (+1.61) and government (+0.90) are what carried the print above zero.
| Component | Contribution to 3.4% (points) | Growth rate |
|---|---|---|
| Consumer spending | +2.52 | 3.8% (AP) |
| Business and housing investment | +1.66 | Nonresidential 12.7%, housing −3.3% (AP); equipment 20.9%, intellectual property 8.4%, structures 5.4% (Trading Economics) |
| Private inventories | +1.61 | n/a |
| Government | +0.90 | 5.3% (Trading Economics) |
| Imports | −2.64 | +19.7% |
| Exports | −0.68 | −5.7% |
How much of the 3.4% is inventories and oil accounting?
Roughly two and a half points, on the tabulated figures. Inventories added 1.61 points, and BEA says the largest contributors were wholesale trade and manufacturing, both led by nondurable goods industries. The inventory build was not stockpiled servers. It was nondurables, and the Q4 2025 revisions show how far this line can move: BEA's third estimate cut growth to 0.5% from 0.7% on weaker wholesale inventories once updated Census data arrived.
Government added 0.90 points, and here the mechanism is unusual. BEA says the government contribution was more than accounted for by federal nondefense consumption, and that the pattern of nondefense spending primarily reflected sales of crude oil from the Strategic Petroleum Reserve. BEA books SPR oil sales as negative nondefense consumption when the oil is sold. If sales were smaller in Q3 than in Q2, the negative shrinks, and measured government spending rises. No new money was spent; the drag simply eased. It should not be read as fiscal stimulus.
Strip those two lines out and the rest of the decomposition sums to about 0.9 points: consumer spending (+2.52) plus fixed investment (+1.66) minus the 3.32-point trade drag. That is roughly what the headline would have shown without the inventory build and the SPR effect. BEA's cleaner gauge ignores trade: real final sales to private domestic purchasers grew 4.9% in Q3 after 4.6% in Q2. Private demand ran fast in both quarters, and the headline swung on inventories, oil accounting and imports.
- Inventories: +1.61 points, led by wholesale and manufacturing nondurables (BEA)
- Government: +0.90 points, driven by the SPR sales pattern in federal nondefense consumption (BEA)
- Final sales to private domestic purchasers: 4.9% vs 4.6% in Q2 (BEA)
How far do advance estimates usually move?
Far enough to change the sentence. BEA publishes three estimates for each quarter. In the last five quarters the gap between the first and third ranged from 0.1 point (Q1 2026, Q3 2025) to 0.9 point (Q4 2025). Q2 2026 opened at 1.5% on 30 July, stayed at 1.5% in the second estimate, then rose to 2.2% in the third on upward revisions to investment, consumer spending and government. Q4 2025 went the other way: 1.4% at the advance, 0.7% at the second, 0.5% at the third, the last cut from weaker wholesale inventories once updated Census data arrived.
Q1 2026 opened at 2.0%, moved down to 1.6% and then back up to 2.1%, with the final leg driven by a downward revision to imports, which are a subtraction in the calculation. Q2 2025 climbed from 3.0% to 3.3% to 3.8%, the last step from consumer spending. Q3 2025 had no normal cycle at all: the autumn 2025 shutdown led BEA to publish a combined 'initial' estimate of 4.3% on 23 December 2025 and an 'updated' 4.4% on 22 January 2026.
Two things stand out for this quarter. First, the lines that drove the biggest recent revisions, imports and wholesale inventories, are exactly the lines doing the heavy lifting in the 3.4% print. Second, revisions do not stop at the third estimate. BEA's September 2026 annual update later lifted Q1 2026 to 2.5% from the 2.1% third estimate, and the September 2025 update cut Q1 2025 to −0.6%.
One group of readers will not care: Kalshi and Polymarket contracts on Q3 2026 GDP settled on 29 October using only the advance estimate, with every 'above' bracket through 2.5% resolving YES. Later revisions are excluded under the market rules. When you see 'markets said 3.4%', that is what it means.
| Quarter | Advance | Second | Third | Main driver of the last revision |
|---|---|---|---|---|
| Q2 2025 | 3.0% | 3.3% | 3.8% | Consumer spending revised up |
| Q3 2025 | 4.3% (initial, shutdown-combined) | n/a | 4.4% (updated) | Shutdown replaced the normal cycle |
| Q4 2025 | 1.4% | 0.7% | 0.5% | Weaker wholesale inventories |
| Q1 2026 | 2.0% | 1.6% | 2.1% | Imports revised down |
| Q2 2026 | 1.5% | 1.5% | 2.2% | Investment, consumer and government revised up |
What does the price side of the same release say?
The GDP report carries its own inflation figures, and they cooled sharply. The price index for gross domestic purchases rose at a 3.0% annual rate in Q3, down from 5.6% in Q2. PCE prices rose 2.1% and core PCE 2.5%. For comparison, the Q2 advance estimate in July had shown quarterly PCE prices at 5.1%, core at 3.4% and gross domestic purchases prices at 5.7%. Current-dollar GDP, which combines real growth and prices, rose 5.6%.
The monthly figures released the same morning tell a different story over a different window. BEA's September Personal Income and Outlays report put headline PCE inflation at 0.5% for the month and 3.6% over the year, with core at 0.3% and 3.1%. The quarterly 2.1% is an annualised three-month rate; the 3.6% is a twelve-month comparison. Both are correct, and neither should be swapped for the other.
September's household numbers were soft underneath the growth story. Personal income rose 0.2%, nominal spending 0.5%, but real spending was flat at less than 0.1%, real disposable income fell 0.3% and the saving rate was 3.8%. The Federal Reserve, meeting the day before the GDP release, held the funds rate at 3.75% to 4% on 28 October. Its statement described activity as 'expanding at a strong pace', an upgrade from 'solid' in September, and said inflation remains elevated. Governor Lisa Cook and Cleveland Fed President Beth Hammack dissented in favour of a quarter-point increase.
- Quarterly GDP measures (Q3, annualised): gross domestic purchases prices 3.0%, PCE 2.1%, core PCE 2.5%
- Monthly September measures: PCE +0.5% m/m and 3.6% y/y, core +0.3% m/m and 3.1% y/y
- FOMC, 28 Oct: rate held at 3.75% to 4%, two dissents for a hike
What to watch next, and what is not confirmed
The next scheduled checkpoints are BEA's second estimate for Q3 on 25 November 2026 at 8:30 a.m. EST, and the FOMC decision on 9 December 2026. The Atlanta Fed's first GDPNow reading for Q4 2026, published on 29 October, was 3.2%. That is a model output, not a forecast of the print, and it will change as data arrive.
Several figures in this piece rest on secondary tabulations rather than BEA's own text. The contribution lines in the table above come from StockTitan citing BEA tables; the equipment rate of 20.9% and private investment at 19.4% are from Trading Economics. We have used the StockTitan tabulation because its five lines reconcile to the 3.4% headline. Check them against Table 2 of BEA's release PDF before reusing them.
BEA's own reliability study is widely cited as showing average revisions of 0.4 point from advance to second and 0.7 to third, without regard to sign. We have not confirmed those figures against the document itself. A reported data-centre construction figure for August appears only on a secondary site and is not in the Census construction release, so it is left out.
- 25 Nov 2026: BEA second estimate for Q3
- 9 Dec 2026: FOMC decision
- Unverified here: BEA's published average-revision figures (0.4 pt advance-to-second, 0.7 pt advance-to-third as cited in summaries), data-centre construction totals
FAQ
How fast did the US economy grow in the third quarter of 2026?
BEA's advance estimate, released 29 October 2026, put real GDP growth at a 3.4% annual rate, up from 2.2% in the second quarter. The figure will be revised in BEA's second estimate on 25 November 2026, with a third estimate to follow.
Was Q3 growth driven by AI investment?
Information processing equipment rose 37.4% and BEA says computers and peripherals led the gain. BEA also says non-automotive capital goods were the largest contributor to the 19.7% rise in imports, and imports subtracted 2.64 points, so the net effect on the headline was far smaller than the gross investment figure.
How much do advance GDP estimates get revised?
In the last five quarters the gap between BEA's advance and third estimates ranged from 0.1 to 0.9 points. Q2 2026 rose from 1.5% to 2.2%; Q4 2025 fell from 1.4% to 0.5%. Annual updates can move them again.
What did the Fed do the day before the GDP report?
On 28 October 2026 the FOMC held the federal funds rate at 3.75% to 4%, described activity as expanding at a strong pace and inflation as elevated. Cook and Hammack dissented in favour of a quarter-point hike. The next decision is 9 December.
What was US PCE inflation in September 2026?
BEA reported headline PCE prices up 0.5% on the month and 3.6% over the year, with core PCE up 0.3% and 3.1% over the year.
Sources
- BEA – GDP (Advance Estimate), 3rd Quarter 2026
- BEA – release PDF with Tables 1–2 (BEA 26-45)
- BEA – Personal Income and Outlays, September 2026
- BEA – GDP Third Estimate, 2nd Quarter 2026
- Federal Reserve – FOMC statement 28 Oct 2026
- Atlanta Fed – GDPNow commentaries
- AP via PBS NewsHour
- StockTitan – Q3 2026 GDP advance estimate tabulation
Figures change. Check the primary sources above.