US real GDP grew at a 3.4% annual rate in the third quarter of 2026, up from 2.2% in the second, according to the Bureau of Economic Analysis’ advance estimate released on 29 October. BEA data published with the release put private investment in data centers at $83.0 billion at an annual rate, the highest since the series began in 2020, against $66.1 billion in the second quarter and $49.2 billion a year earlier.
Information processing equipment and software together added 1.09 percentage points to the quarter’s growth, by our arithmetic from BEA’s contribution table: about 32% of the 3.4%.
Short explainer · US GDP and data centers
A vertical news Short narrated from this piece’s headline and opening line.
Business Pill · Capex and Depreciation
An evergreen Business Engineer explainer on capital expenditure and depreciation. It explains the general idea only and does not discuss this GDP release.
The key insight: By our arithmetic from BEA’s contribution table, information processing equipment and software added 1.09 of the third quarter’s 3.4 points of growth, about 32%, against about 24% in the second quarter and about 10% a year earlier.
BEA says the increase reflected higher investment, consumer spending and government spending, partly offset by lower exports, while imports, which are subtracted in the calculation of GDP, rose. It says the increase in investment “primarily reflected increases in private inventory investment and in equipment”.
Within equipment, BEA says, “the increase was led by information processing equipment (mainly computers and peripheral equipment)”, based primarily on Census Bureau import data. Real final sales to private domestic purchasers grew 4.9%, against 4.6% in the second quarter, and the PCE price index rose 2.1%.
Contribution of information processing equipment plus software to US real GDP growth, percentage points at seasonally adjusted annual rates: 0.38 (Q3 2025), 0.72 (Q4 2025), 1.39 (Q1 2026), 0.52 (Q2 2026), 1.09 (Q3 2026), against GDP growth of 3.9%, 0.2%, 2.5%, 2.2% and 3.4%. The sum of the two BEA lines is our arithmetic. Source: US Bureau of Economic Analysis, Tables 1.5.2 and 1.1.1, Q3 2026 advance estimate.
What Was Posted on X
MACRO DATA JUST DROPPED
โข PCE 3.6% YoY vs. Est. 3.6% โข Core PCE 3.1% YoY vs. Est 3.1%
Wall St Engine on X, 29 Oct, quoting White House senior adviser Kevin Hassett: if US October inflation is really high, the Fed is probably to act.
The Lines Closest to AI
BEA’s contribution table, Table 1.5.2, gives information processing equipment 0.86 percentage point of the 3.4% growth, of which computers and peripheral equipment gave 0.72. Software gave 0.23 point. Structures gave 0.15 point, after zero in the second quarter and negative contributions in the three quarters before that.
By our arithmetic, the 1.09 points from information processing equipment and software compare with 0.52 point in the second quarter, about 24% of that quarter’s 2.2% growth, and 0.38 point a year earlier, about 10% of 3.9%. In the first quarter of 2026 the two lines gave 1.39 points of 2.5% growth.
Spending on data-center structures is not broken out in the contribution table. In current dollars, BEA’s Table 5.4.5 shows it at $82,979 million at an annual rate, up about 25.5% from the second quarter and about 68.7% from a year earlier, by our arithmetic. The first quarter of 2020, the earliest in the series, was $9,524 million.
The Import Offset
Imports subtracted 2.64 points from growth and exports a further 0.68, so net exports subtracted 3.32 points. BEA says “The largest contributor to imports was capital goods, except automotive”. Its import table gives computers, peripherals and parts 8.38 points of the 19.7% rise in real imports.
By our rough arithmetic, that share of import growth corresponds to about 1.1 points of the GDP drag from imports, close to the 1.09 points that information processing equipment and software added. BEA does not net these lines against each other, and imported computers and parts also go to uses other than business investment.
How It Compares With the Previews
A Reuters preview published before the release was headlined “AI-related investment, consumers likely boosted US economic growth in third quarter”. An IndexBox headline the same morning put expected growth at 3.2%. On 28 October a Reuters headline said AI-related demand pushed imports to a record high as the US goods trade deficit widened.
ING wrote in August that “We believe AI accounts for a third of economic growth in 2026”, as we reported on 28 October. ING’s figure is a full-year estimate with its own definition of AI; the 32% here is one quarter’s contribution from two BEA lines, so the two measure different things.
The Structural Read
As we read the release, the quarter’s growth leaned on investment: BEA names inventories and equipment, and within equipment, computers and peripheral equipment.
By our arithmetic, private investment in data-center structures rose about 68.7% in a year in current dollars, to an $83.0 billion annual rate, while structures as a whole contributed 0.15 point to real growth.
As we read the import table, much of the hardware came from abroad: computers, peripherals and parts gave 8.38 points of the 19.7% rise in real imports, and imports subtracted 2.64 points from GDP growth.
US Bureau of Economic Analysis, 29 October 2026
“Within equipment, the increase was led by information processing equipment (mainly computers and peripheral equipment)”
Three Implications
FOR HYPERSCALER CAPEX By our arithmetic, the BEA lines that include servers and software added 1.09 points to third-quarter growth; BEA does not say which buyers drove them.
FOR THE TRADE BALANCE BEA says the largest contributor to imports was capital goods, except automotive; net exports subtracted 3.32 points from growth.
FOR THE NEXT ESTIMATE BEA says the second estimate, with corporate profits, is due on 25 November; advance figures can be revised.
The framework puts it this way: “The economy is only one layer of the story.”
As we read it, the BEA release shows the build reaching the national accounts through computers, software and data-center structures. The framework does not discuss this release or the third quarter of 2026.
BEA does not publish an artificial-intelligence category; information processing equipment, software and data-center structures also include spending unrelated to AI. The data-center figure is in current dollars and is not a contribution to real growth.
This is an advance estimate. BEA says the next estimate, with corporate profits, is due on 25 November, and its figures can be revised. The import offset above is a rough calculation of ours, not a BEA figure.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
BEA’s advance estimate puts third-quarter growth at 3.4%. Information processing equipment and software added 1.09 points, about 32% of it by our arithmetic, and data-center investment reached a record $83.0 billion annual rate, while imports led by capital goods subtracted 2.64 points.
A note on sourcing. Figures and quotations come from the US Bureau of Economic Analysis advance estimate of third-quarter 2026 GDP, released on 29 October 2026, and BEA’s published NIPA data (Tables 1.1.1, 1.5.2, 4.2.2 and 5.4.5). Preview figures are attributed to the outlets’ headlines. Sums, shares, growth rates and the rough import offset are our arithmetic from BEA’s figures.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
