TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
The US economy grew at a 3.4% annual rate in the third quarter of 2026, and computers and software contributed to that expansion. The figures show exposure to AI-related investment, but they do not establish how much spending was specifically for AI or predict what a slowdown would do.
The US economy grew at a 3.4% annual rate from July through September, according to the Bureau of Economic Analysis’ advance estimate published October 29, while business spending on computers and peripheral equipment rose at a 66.7% annual rate. The figures show that technology investment contributed to growth, but they do not isolate spending on artificial intelligence or establish how the economy would perform if that spending eased.
Consumer spending added 2.52 percentage points to the quarter’s growth, the largest contribution among the categories listed in the source material. Spending grew at a 3.8% annual rate, with increases in health care, restaurants and hotels, and prescription drugs. Equipment investment added 1.16 points, while intellectual property products, including software and research, added 0.47 points. The BEA’s tables show that information processing equipment contributed 0.86 points and software 0.23 points: together, 1.09 points, or about 32% of the 3.4% growth rate.
Those figures cover technology categories that can include AI-related purchases, such as servers and chips, but the national accounts do not identify what share was bought specifically for AI. The source author calculates that information processing equipment and software averaged about 45% of growth over the previous four quarters, based on BEA Table 1.5.2. That is the author’s calculation, not a BEA estimate of AI’s contribution.
Imports subtracted 2.64 percentage points from third-quarter growth, with capital goods the largest contributor. Imports rose at a 19.7% annual rate. The source author estimates that imported computer equipment reduced growth by about 1.1 points, compared with a 0.72-point contribution from business purchases of computers. That estimate is also the author’s calculation from BEA data. The comparison illustrates why domestic investment and total production in the United States are not the same measure: imported equipment is not US production and is subtracted in GDP accounting.
Other categories also shaped the result. Inventories added 1.61 points, government spending 0.90 points, and exports subtracted 0.68 points. Housing investment subtracted 0.12 points. The figures are contributions to the annualized growth rate and may not sum exactly because of rounding.
What Happens to US Growth If the AI Boom Loses Steam?
Technology investment helped power a strong quarter. The GDP data show exposure to computers and software, but they do not isolate AI spending or predict the impact of a slowdown.
percentage points
percentage points
equipment + software points
percentage points
Growth came from several directions
The reported 3.4% is an annualized pace: the quarter’s actual output increase was 0.8%. Growth followed 2.2% in Q2 and exceeded the 3% forecast cited from a FactSet survey by the Associated Press.
Contributions are percentage points of annualized growth. Bars show relative magnitude; negative contributions extend from the right. Totals may differ because of rounding.
Information processing equipment contributed 0.86 points and software 0.23, together about 32% of the 3.4% growth rate. These categories can include AI-related purchases, but the national accounts do not identify what share was specifically for AI.
Relevant categories are not an AI total
Computers, peripheral equipment and software serve many business uses. The source author’s calculations highlight exposure in BEA tables; they are not official estimates of AI’s contribution.
Information processing
Added 0.86 percentage points to Q3 growth. Purchases can include servers and chips used in AI infrastructure, alongside other equipment.
Software
Added 0.23 points. The category captures software investment broadly, without separating AI products from other software.
About 45% average
The author estimates these categories averaged about 45% of growth over the prior four quarters. This is a calculation from BEA Table 1.5.2, not a BEA AI estimate.
Imports subtracted 2.64 points as they rose at a 19.7% annual rate. The source author estimates imported computer equipment reduced growth by about 1.1 points, while domestic business computer purchases contributed 0.72 points. Imported equipment is subtracted because GDP measures production in the United States.
Effects may be concentrated before they show up in headline growth
A slower build-out could reach connected suppliers and projects. The source material does not quantify those exposures or establish how much current activity depends on AI demand.
Sites & suppliers
Data centers use computers, construction, electricity and financing. Fewer projects could affect equipment makers, contractors, power projects and lenders.
Jobs & local activity
Construction supports work at building sites and specialized suppliers. The cited September report added 29,000 jobs, below the cited prior-year monthly average of 45,000; one month does not establish an AI-related employment trend.
Investment exposure
The source argues stock-heavy households could be affected if build-out companies weaken. It provides no market estimate or measure of the risk.
Private data-center construction ran at an $85.0 billion annual rate in August, up 73% year over year. Factory construction was down 20%. These measures are not direct readings of AI spending or forecasts.
The supplied account says the Nasdaq fell 2.9% on Friday, October 30, after TechCrunch reported “cracks are emerging in the AI spending boom.” One day does not confirm a lasting investment slowdown.
GDP records a net result, not a single AI line
| What the data show | What they do not establish |
|---|---|
| Broad investment in computers, peripheral equipment and software | What share was purchased specifically for AI |
| Imports subtracted 2.64 points; capital goods were a major factor | How lower server imports would compare with lost domestic investment |
| Inventories added 1.61 points to Q3 growth | Whether stock-building will persist or reflect future demand |
| August construction comparisons and one monthly jobs report | Whether later building or employment changes are connected to AI |
| Strong quarterly growth alongside substantial technology investment | A quantified GDP, jobs or market scenario if AI spending slows |
The advance estimate is the BEA’s first of three estimates and relies partly on assumptions where September data were not yet available. GDP measures output produced in the United States; imports are subtracted and exports count as domestic production sold abroad.
The next estimate adds evidence—not a verdict
The BEA’s second estimate is due November 25, 2026. It will incorporate additional information and may revise both the growth rate and its components.
GDP revision
Compare updated equipment, software, imports and inventory estimates.
Company plans
Look for later investment plans tied to data centers and technology.
Build & trade
Track construction figures and trade data for changes in the cycle.
Employment
Use later reports to assess whether labor-market patterns persist.
The quarter combined 3.4% annualized growth with substantial technology investment and imports. The scale of AI-specific spending—and the consequences of any slowdown—remains unmeasured.
Where a Pullback Could Land
A slowdown in AI-related investment could affect the suppliers and projects most directly tied to the build-out, even if its effect on the national growth rate were smaller than headlines suggest. Data centers require computers, construction, electricity and financing; a reduction in new projects could therefore reach equipment makers, contractors, power projects and lenders. The source material does not quantify those exposures or show how much of current activity depends on AI demand.
The import calculation complicates the picture. If businesses bought fewer imported servers, imports could also rise less or fall, which would reduce the negative contribution from imports in GDP arithmetic. That would not automatically replace lost domestic construction, services or equipment investment, and the overall effect would depend on what spending changed and where the equipment was produced. The national accounts record the net result across these categories, rather than a single AI line.
Workers and households could also experience uneven effects. Data-center construction supports activity at building sites, while a downturn could affect contractors and specialized suppliers in places with concentrated projects. The September employment report cited in the source showed 29,000 jobs added, below the cited prior-year monthly average of 45,000. That single month does not establish that AI investment has failed to create jobs or that a spending slowdown would cause broad job losses.
Investors may face another channel of exposure: the source argues that households with stock-heavy savings could be affected if companies tied to the build-out weaken. The material does not provide market estimates or identify the size of that risk. The practical question is whether investment continues to support output and earnings, and how concentrated any pullback would be.
Products Worth Considering
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
Bose TV Speaker Compact Soundbar with Bluetooth and HDMI-ARC, Black
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
As an affiliate, we earn on qualifying purchases.
How the Quarter Added Up
The reported 3.4% is an annualized rate: it expresses what the third-quarter pace would amount to over a year if it continued. The actual increase in output within the quarter was 0.8%. The result followed 2.2% annualized growth in the second quarter and exceeded the 3% forecast cited from a FactSet survey by the Associated Press.
The advance estimate is the first of three BEA estimates. It relies partly on assumptions where September data were not yet available. It reports output produced in the United States, not simply the spending by US households and businesses. For that reason, imports are subtracted, while exports are included as production sold abroad.
Separate construction data provide another view of the investment cycle. Private data-center construction ran at an annual rate of $85.0 billion in August, up 73% from a year earlier, according to the Census Bureau figures cited in the source. Factory construction was down 20% over the same period. These are construction measures, not a direct reading of AI spending or a forecast of future building.
The source describes the stock-market mood as shifting after TechCrunch reported that “cracks are emerging in the AI spending boom.” The Nasdaq fell 2.9% on Friday, October 30, according to the supplied account. A one-day index decline does not by itself confirm a lasting slowdown in investment.
What the GDP Tables Cannot Show
The BEA categories do not isolate artificial intelligence. Computers, peripheral equipment and software can serve many business purposes, so their contributions cannot be treated as a precise measure of AI investment. The source author’s estimates identify relevant categories and make calculations from BEA tables, but they are not official AI-specific GDP figures.
The quarter’s growth also included a large inventory contribution. The supplied figures do not establish whether that stock-building will persist, or how much it reflects expectations about future demand. Nor do they show whether the September employment slowdown was connected to AI investment.
The data-center and factory construction figures compare August with the same month a year earlier; they do not establish the direction of building activity after August. The source material also offers no quantified scenario for how much GDP, employment or market value would change if AI-related spending slowed. Those consequences remain uncertain.
The November GDP Revision
The BEA’s second estimate is due November 25, 2026. It will incorporate additional information and may revise the third-quarter growth rate and its components. Readers can compare the updated estimates for equipment, software, imports and inventories to see whether the advance report’s pattern holds.
Further evidence would be needed to judge whether the AI build-out is losing momentum: later company investment plans, construction figures, trade data and employment reports could each add detail. None alone would settle the question. For now, the reported quarter shows strong growth alongside substantial technology investment and imports, while the scale and economic consequences of any slowdown remain unmeasured.
Key Questions
How fast did the US economy grow in the third quarter of 2026?
The BEA’s advance estimate put real GDP growth at a 3.4% annual rate from July through September. The actual increase within the quarter was 0.8%.
How much did computers and software contribute?
Information processing equipment contributed 0.86 percentage points and software 0.23 points, for a combined 1.09 points of the 3.4% annualized growth rate. These categories include technology purchases but do not identify how much was specifically for AI.
Does the report show that AI investment is slowing?
No. The supplied figures show fast growth in business spending on computers and peripheral equipment, alongside concerns reported by TechCrunch and a Nasdaq decline on Friday. They do not establish a broad or sustained drop in AI investment.
Why did imports reduce measured growth?
GDP measures production in the United States, so goods produced abroad are subtracted from spending totals. Imports reduced third-quarter growth by 2.64 percentage points, with capital goods the largest contributor.
When will the next GDP estimate be released?
The BEA’s second estimate is scheduled for November 25, 2026. It may revise the advance estimate as more data become available.
Source: Thorsten Meyer AI
