AI in Construction Statistics 2026: Adoption, Growth and Uses

AI in construction statistics: what share of construction companies use AI in 2026?

17.2% of U.S. construction businesses used AI in the latest Census reading, against 23.8% of all businesses. That is the highest construction reading since the Census Bureau revised its AI question in November 2025.

17.2% of construction businesses told the Census Bureau they used artificial intelligence in any business function in the two weeks from September 7, 2026 to September 20, 2026. The figure comes from the Business Trends and Outlook Survey (BTOS) period collected September 21, 2026 to October 4, 2026 and released October 8, 2026.

The all-sector rate in the same period was 23.8%, so construction trails by 6.6 percentage points. The construction estimate carries a standard error of 0.72 points, so single biweekly readings move around. Averaged over the six latest periods (July 13 to October 4, 2026), construction stands at 15.2% and all sectors at 22.8%.

AI use (BTOS): A business counts as an AI user if it answered Yes to: 'In the last two weeks, did this business use Artificial Intelligence (AI) in any of its business functions?' The questionnaire gives machine learning, natural language processing, virtual agents and voice recognition as examples. Shares are firm-weighted: a five-person subcontractor counts the same as a national builder.

How many construction companies use AI?

About 135,000 U.S. construction employer firms used AI in the latest two-week reading, on a CB&C estimate that applies the Census share to the 2022 firm count.

Roughly 135,000 construction employer firms used AI in the two weeks to September 20, 2026. The estimate multiplies the latest BTOS construction share (17.2%) by the 782,487 construction employer firms in Statistics of U.S. Businesses for 2022, the latest year published.

Using the six-period average (15.2%) gives about 119,000 firms. About 159,000 firms expect to use AI within six months, based on the 20.3% expected-use share. These are order-of-magnitude figures: the survey share is from 2026, the firm count is from 2022, and Census tabulates firms that operate in several sectors separately from single-sector construction firms.

An earlier Census benchmark, covering 2020 to 2022, is far lower. The 2023 Annual Business Survey found 3,384 construction firms, 0.9% of those answering, used AI as part of their processes or methods in 2020 to 2022, against 3.3% of all firms.

Is construction behind other industries in AI adoption?

Yes. Construction ranks 14th of 19 sectors on AI use, with 5 sectors below it. The gap to the all-sector rate has narrowed under both versions of the Census question.

Construction ranks 14th of 19 sectors on mean AI use over the six latest Census periods, at 15.2%. Information leads at 45.9%, followed by professional, scientific and technical services (41.9%) and finance and insurance (38.3%). Retail trade (15.8%) sits just above construction; other services (12.8%), transportation (10.6%), agriculture, accommodation and food, and mining (8.3%) sit below.

The relative gap is shrinking. Under the revised question, the construction rate was 0.43 of the all-sector rate in November 2025 and 0.72 in the latest period. Under the original question it moved from 0.27 in September 2023 to 0.45 in September 2025. In absolute terms the original-question gap widened, from 2.7 to 5.5 points, because all-sector use grew faster from a higher base.

The ranking is not new. A Census working paper covering September 2023 to February 2024 put construction at 1.4% current use, tied with agriculture as the lowest sector, against 18.1% in information. In the 2023 Annual Business Survey (data year 2022), construction was the second-lowest of 19 sectors at 0.9% of firms, 0.27 of the all-sector rate.

How fast is AI adoption growing in construction?

Construction AI use rose 65% between the first and latest six-period averages of the revised Census series, more than double the 28% rise across all sectors.

Construction AI use averaged 9.2% in the six periods collected November 17, 2025 to February 8, 2026, and 15.2% in the six periods collected July 13 to October 4, 2026. That is a rise of 6.0 points, or 65%. All sectors rose 5.0 points, or 28%, over the same windows.

Point to point, construction went from 7.4% to 17.2% in under a year, 2.3 times the starting level and a gain of 9.8 points, while the all-sector rate gained 6.5 points. Endpoints are single biweekly estimates, so the six-period averages are the steadier measure.

The original question shows the same direction over a longer span. Construction AI use rose from 1.0% in the period collected September 11 to 24, 2023 to 4.5% in the period collected September 22 to October 5, 2025, 4.5 times the starting level. The all-sector rate rose 2.7 times, from 3.7% to 10.0%.

Whether wider AI use is lifting output per worker on job sites is a separate question, tracked in our construction productivity research.

How many construction firms expect to use AI in the next six months?

20.3% of construction businesses expect to use AI within six months, 3.1 points above current use. Recent construction use has run ahead of what firms expected.

20.3% of construction businesses said they expect to use AI in a business function during the next six months, in the period collected September 21, 2026 to October 4, 2026. The all-sector figure was 27.7%. Averaged over six periods, expected use is 17.4% in construction and 26.5% across all sectors.

The expected-minus-current gap is smaller in construction (2.2 points on six-period averages) than in all sectors (3.7 points). Taken alone, that would point to slower growth ahead.

Construction firms have under-predicted their own uptake, though. Expected use averaged 13.1% in the six periods collected January 12 to April 5, 2026. Actual use six months later averaged 15.2%, 2.1 points higher. For all sectors the shortfall was 0.8 points (22.0% expected, 22.8% actual).

What is AI used for in construction?

Mostly office work. Construction firms use AI most in sales and marketing, strategy and IT; only 0.7% used it in production work, the category that includes construction itself.

Sales and marketing is the most common AI function in construction: 7.7% of all construction businesses used AI there in the six months to early 2026. Strategy and business development (7.5%), information technology (6.9%) and research and development (6.3%) follow, according to the Census 2026 AI supplement, collected November 17, 2025 to February 8, 2026.

Production of goods, which the questionnaire illustrates with 'making or assembling products, construction', drew AI use from 0.7% of construction firms, against 1.7% of all firms. Quality management (1.6%) and distribution (0.5%) were also rare. Construction comes closest to the all-sector rate in sourcing and purchasing (a ratio of 0.88) and human resources (0.79).

Generative AI use follows the same office pattern. 12.3% of construction firms said employees used generative AI for work tasks. Among those firms, 86.2% used it for writing or editing, 48.8% for searching for information, 40.9% for interpreting or summarizing documents and 37.6% for paperwork or filing.

Industry surveys point the same way. In the AGC and Sage outlook survey of member firms (November to December 2025), 45% of firms deployed AI for office and administrative functions, 23% for estimating and 20% for design or preconstruction. Those are the paperwork-heavy stages of the commercial construction process, not field installation.

Earlier Census data on motives match. Among construction firms that used AI in 2020 to 2022, 44.1% adopted it to improve the quality or reliability of processes and 27.6% to automate tasks performed by labor (2023 Annual Business Survey).

Which construction firms use AI most?

Large firms and building contractors. Construction firms with 250 or more employees averaged 46.6% AI use, 3.4 times the rate for firms with 1 to 4 employees.

AI use in construction climbs with firm size. Over the six latest periods, firms with 1 to 4 employees averaged 13.7%, firms with 20 to 49 employees 21.6%, firms with 100 to 249 employees 33.9% and firms with 250 or more 46.6%.

By subsector, construction of buildings (NAICS 236, general contractors and builders) averaged 20.2%. Heavy and civil engineering construction (NAICS 237) averaged 12.4% and specialty trade contractors (NAICS 238), the largest group of firms, 12.8%. Because most construction firms are small trade contractors, the sector rate sits near the low end.

Subsector and size estimates rest on fewer responses than the sector total, and their biweekly values swing widely. The latest single reading for building construction was 25.4%, well above its six-period mean.

Will AI replace construction jobs?

Federal data show little sign of it so far. 1.1% of construction AI users said AI cut their total employment, and BLS projects construction and extraction occupations to grow 5.1% by 2035.

1.1% of construction firms that used AI said it decreased their total employment in the six months to early 2026, while 1.5% said it increased employment and 97.4% saw no change. The all-sector shares were 2.0% and 2.3%.

Construction AI users mostly add AI to existing work. 32.4% used it to supplement or enhance a task done by an employee, while 5.0% used it to perform a task an employee used to do (all sectors: 10.1%).

Older Census data cover a longer window. Among construction firms that adopted AI in 2020 to 2022, 7.8% said worker numbers fell, 14.3% said they rose and 77.9% reported no change (2023 Annual Business Survey).

BLS projects construction and extraction occupations to grow 5.1% from 2025 to 2035, against 3.5% for all occupations, and construction industry employment to grow 5.0%. BLS projects office and administrative support jobs to fall 4.0%, citing automation tools including those powered by AI. BLS states that the projections assume technological progress in line with historical patterns, so they are not a forecast of AI effects.

BLS also rates occupations by relative AI exposure. 66.8% of 2025 construction and extraction employment is in 'Low' exposure occupations, against 17.8% of all employment. Construction laborers and carpenters are rated Low. Construction managers and cost estimators are rated High: BLS projects construction managers to grow 9.1% and cost estimators to decline 3.1% from 2025 to 2035. BLS notes that exposure does not imply job loss. Hiring pressure in the trades is covered in our construction labor shortage report, and pay and entry routes in construction careers.

Why are most construction firms not using AI?

Most say it does not apply to them. Among construction firms with no plans to use AI, 69.0% gave that reason, more than the 61.6% across all sectors.

69.0% of construction businesses not planning to use AI said it is not applicable to their business (2026 AI supplement). The next most common reasons were a lack of knowledge of what AI can do (22.4%), privacy or security concerns (18.9%) and a view that AI is not mature enough (10.4%). Cost was cited by 4.9%.

Adopters in construction also invest less in making AI work. 75.0% of construction AI users made no organizational changes to use AI, against 64.3% of all AI users. 10.0% developed new workflows and 7.5% trained staff, against 15.4% and 15.0% across all sectors.

In the 2023 Annual Business Survey, 76.9% of construction firms rated AI 'not applicable' to their processes in 2020 to 2022, against 64.4% of all firms.

Why do industry surveys report higher AI adoption in construction?

They ask different questions of different samples. The AGC figure of 61% combines current use with plans to invest, and comes from member firms that skew larger than the industry.

The most cited industry figure, 61% from the AGC and Sage 2026 outlook, counts firms that 'use AI or plan to increase investments in it', up from 44% a year earlier. It surveyed 951 AGC members from November 4 to December 15, 2025; only 17% had fewer than 20 employees. Census BTOS weights responses to represent all employer businesses and asks only about use in the last two weeks. In the BTOS periods collected during those weeks, construction use was 7.4% and 9.3%.

• AGC and Sage (U.S. member firms, 951 respondents, November to December 2025): 61% use AI or plan to increase AI investment; 45% use AI for office and administrative functions.
• RICS (global, more than 2200 professionals, Q1 2025, 10% from the Americas): 45% reported no AI implementation, 34% early pilots, just under 12% regular use in specific processes, 1.5% use across multiple processes.
• Dodge Construction Network and CMiC (U.S., 235 contractors, September and October 2025, as reported by Construction Dive): 85% expect to spend less time on repetitive tasks; 57% cite unreliable or inaccurate AI output as a concern.
• Census BTOS (U.S., weighted to all employer businesses, biweekly): 17.2% of construction businesses used AI in the two weeks to September 20, 2026.

Market-size estimates for AI in construction disagree even more. For 2025, Precedence Research puts the global market at $1.63 billion, Global Market Insights at $3.9 billion and Mordor Intelligence at $11.1 billion, a 6.8-fold spread. Forecasts run to $24.7 billion by 2035 (Precedence), $25.6 billion by 2035 (Global Market Insights) and $27.92 billion by 2031 (Mordor). These are vendor estimates with methods that are not published on the pages cited.

Why did Census AI figures for construction jump in late 2025?

Census changed the question. Construction read 4.5% in September 2025 on the old wording and 7.4% in November 2025 on the new one; part of that step is a measurement change, not adoption.

From November 17, 2025, BTOS asks whether a business used AI 'in any of its business functions' instead of 'in producing goods or services'. A Census working paper estimates that the old concept, rebuilt from the new supplement's production and service-provision items, would read about 6% across all sectors, below the last pre-revision estimate of 10%, so the old question already captured some uses beyond production.

No data exist for October 6 to November 16, 2025: the federal funding lapse stopped collection, and Census does not backfill those periods.

Response is thin in construction. The 2026 AI supplement had a 6.0% unit response rate in construction (9,400 responses), the second lowest of 19 sectors, against 10.0% nationally. Census adjusts weights for nonresponse, but low response raises the risk that responding firms differ from the rest.

The numbers in one list

1. 17.2% of U.S. construction businesses used AI in any business function in the two weeks to September 20, 2026 (Census BTOS, collected September 21, 2026 to October 4, 2026).^[2]
2. 23.8% of all U.S. businesses used AI in the same period, leaving construction 6.6 percentage points behind.^[1][2]
3. Averaged over the six latest biweekly periods (July 13 to October 4, 2026), construction AI use was 15.2% vs 22.8% for all sectors.^[1][2]
4. Construction ranks 14th of 19 U.S. sectors on AI use; Information leads at 45.9%, 3.0 times the construction rate.^[2]
5. Construction AI use rose from a 9.2% average in November 2025 to February 2026 to 15.2% in July to October 2026, a 65% increase vs 28% for all sectors.^[1][2]
6. The construction gap to the all-sector rate narrowed from 9.9 points in November 2025 to 6.6 points in the period collected September 21, 2026 to October 4, 2026.^[1][2]
7. Under the original Census question, construction AI use rose from 1.0% in September 2023 to 4.5% in September 2025, a 4.5-fold increase.^[3]
8. 20.3% of construction businesses expect to use AI within six months (period collected September 21, 2026 to October 4, 2026), 3.1 points above current use.^[2]
9. Only 0.7% of construction businesses used AI in production work in the six months to early 2026; sales and marketing (7.7%) and strategy (7.5%) led.^[4]
10. Construction firms with 250 or more employees averaged 46.6% AI use vs 13.7% for firms with 1 to 4 employees (July to October 2026).^[7]
11. Building contractors (NAICS 236) averaged 20.2% AI use, above heavy and civil (12.4%) and specialty trades (12.8%), July to October 2026.^[8]
12. 1.1% of construction AI users said AI reduced their total employment in the six months to early 2026; 97.4% reported no change.^[4]
13. BLS projects construction and extraction occupations to grow 5.1% from 2025 to 2035, faster than the 3.5% for all occupations.^[18]
14. 66.8% of 2025 construction and extraction employment is in occupations BLS rates as 'Low' AI exposure, vs 17.8% of all U.S. employment.^[19]

Calculations

Download the calculations (CSV)

Summary

• 17.2% of U.S. construction businesses used AI in the latest Census reading, against 23.8% of all businesses.
• Construction ranks 14th of 19 sectors on average AI use (15.2%); information leads at 45.9%.
• Construction use is growing faster than the economy: up 65% between late 2025 and late summer 2026, against 28% for all sectors.
• The gap to the all-sector rate narrowed from 9.9 to 6.6 points since November 2025.
• Use is mostly in office functions: 7.7% of construction firms used AI in sales and marketing, 0.7% in production work.
• Large contractors lead: 46.6% of construction firms with 250 or more employees use AI, against 13.7% of firms with 1 to 4.
• Job effects are small so far: 1.1% of construction AI users reported lower employment because of AI, and BLS projects construction and extraction occupations to grow 5.1% by 2035.

Frequently asked questions

Related guides, services and research

• The Commercial Construction Process, From Feasibility to Closeout
• Design-Build Contractors: One Contract for Design and Construction
• Construction Careers: Paths, Pay and Where the Openings Are
• Commercial Construction Spending Tracker: August 2026 Data by Building Type
• Commercial Construction Costs 2026: The CB&C Margin Squeeze Index

Methodology

Source selection. The primary dataset is the Census Bureau's Business Trends and Outlook Survey (BTOS), a biweekly, nationally representative survey of about 1.2 million U.S. employer businesses split into six panels. We used the official download files (National, Sector, Subsector, Sector by Employment Size Class, AI Core Questions and both AI supplement tables) from census.gov/hfp/btos/downloads, opened and saved unmodified on October 8, 2026. Secondary Tier 1 sources are the 2023 Annual Business Survey technology module (data year 2022, the latest technology module), Statistics of U.S. Businesses 2022, BLS Employment Projections 2025 to 2035 and the BLS AI exposure table, and two Census Center for Economic Studies working papers.

Inclusion and exclusion. Only figures read directly from an opened file or page are used. Industry surveys (AGC and Sage, RICS, Dodge via Construction Dive) and market-size estimates are reported as headline figures with sample and method, never blended with Census rates. A market estimate from Fortune Business Insights was excluded because its page could not be opened. Claims from roundup pages that could not be traced to a primary source were excluded (see excluded.md).

Two question wordings. Census asked about AI 'in producing goods or services' from September 2023 to October 5, 2025 and about AI 'in any of its business functions' from November 17, 2025. We report the two series separately and never compute change across the break. No data exist for October 6 to November 16, 2025 because of the federal funding lapse.

Derived metrics. Gaps are all-sector rate minus construction rate in percentage points. Six-period means are simple averages of the published biweekly estimates for periods 202615 to 202620 (one full rotation of the six sample panels); they are CB&C calculations, not Census estimates. Sector ranks use those means for the 19 two-digit NAICS sectors; multi-sector firms (Census code XX) are excluded. Where Census suppressed a value ('S'), the mean uses the periods published (agriculture 3, management of companies 4). Firm counts multiply a BTOS share by the SUSB 2022 construction firm count and are rounded to the nearest 1,000. Every derived value is listed in derived_metrics.csv with its formula and fact-sheet rows.

Limitations. BTOS is an experimental data product. Construction single-period estimates in the revised series have standard errors from 0.44 to 1.15 points, and the construction unit response rate in the 2026 AI supplement was 6%. Rates are firm-weighted, so they describe the share of businesses, not the share of workers or revenue. The ABS (data year 2022) and BTOS use different questions and reference periods and are not comparable as a trend. BLS AI exposure categories are occupation-based and do not measure job loss.

Last update: 2026-10-08, using the BTOS release of October 8, 2026 (period 202620). Next BTOS release: October 22, 2026. Not advice: this report describes published statistics and does not recommend any business, technology or employment decision.

This report presents published data with its limitations. It is not legal, financial or investment advice.

Sources, ranked by quality

Tier 1: primary and official sources (19)

• [1] U.S. Census Bureau. Business Trends and Outlook Survey (BTOS), National response estimates (National.xlsx). Biweekly collection periods, September 2023 to October 4, 2026 (released 2026-10-08). Accessed 2026-10-08.
• [2] U.S. Census Bureau. BTOS, Sector response estimates and standard errors (Sector.xlsx). Biweekly collection periods to October 4, 2026 (released 2026-10-08). Accessed 2026-10-08.
• [3] U.S. Census Bureau. BTOS, AI core questions, original wording (AI Core Questions.xlsx). Collection periods September 11, 2023 to October 5, 2025. Accessed 2026-10-08.
• [4] U.S. Census Bureau. BTOS, 2026 AI supplement tables (AI_Supplement_Table_2026.xlsx). Pooled collection November 17, 2025 to February 8, 2026 (published 2026-04-23). Accessed 2026-10-08.
• [5] U.S. Census Bureau. BTOS, 2026 AI supplement unit response rates (AI26_URR.xlsx). November 17, 2025 to February 8, 2026. Accessed 2026-10-08.
• [6] U.S. Census Bureau. BTOS, first AI supplement tables (AI_Supplement_Table.xlsx). Pooled collection December 4, 2023 to February 25, 2024. Accessed 2026-10-08.
• [7] U.S. Census Bureau. BTOS, Sector by employment size class estimates (Sector by Employment Size Class.xlsx). Biweekly collection periods to October 4, 2026. Accessed 2026-10-08.
• [8] U.S. Census Bureau. BTOS, Subsector (3-digit NAICS) estimates (Subsector.xlsx). Biweekly collection periods to October 4, 2026. Accessed 2026-10-08.
• [9] U.S. Census Bureau. Business Trends and Outlook Survey Methodology, Version 5 (last updated July 9, 2026). Survey design as of July 2026. Accessed 2026-10-08.
• [10] U.S. Census Bureau. Business Trends and Outlook Survey (BTOS) Data (program page). Program description, page revised August 13, 2026. Accessed 2026-10-08.
• [11] U.S. Census Bureau. AI Use at U.S. Businesses: Large Firms With at Least 20 Employees Biggest AI Users (America Counts, May 26, 2026). BTOS, December 14, 2025 to May 3, 2026. Accessed 2026-10-08.
• [12] U.S. Census Bureau. How Many U.S. Businesses Use Artificial Intelligence? (America Counts, November 28, 2023). BTOS, September to November 2023. Accessed 2026-10-08.
• [13] U.S. Census Bureau. How AI and Other Technology Impacted Businesses and Workers (America Counts, September 17, 2025). 2023 Annual Business Survey, data year 2022. Accessed 2026-10-08.
• [14] U.S. Census Bureau, Center for Economic Studies. Bonney et al., Tracking Firm Use of AI in Real Time: A Snapshot from the Business Trends and Outlook Survey, CES Working Paper 24-16R (October 2024). BTOS, September 2023 to February 2024. Accessed 2026-10-08.
• [15] U.S. Census Bureau, Center for Economic Studies. Bonney et al., The Microstructure of AI Diffusion: Evidence from Firms, Business Functions, and Worker Tasks, CES Working Paper 26-25 (April 2026). BTOS 2026 AI supplement, November 2025 to January 2026 reference period. Accessed 2026-10-08.
• [16] U.S. Census Bureau. 2023 Annual Business Survey, Module Business Characteristics, table AB2200MCB01 (technology adoption, data year 2022). Data year 2022 (latest ABS technology module). Accessed 2026-10-08.
• [17] U.S. Census Bureau. Statistics of U.S. Businesses 2022, U.S. firms by sector and enterprise employment size. 2022 (released 2025-04-10; latest SUSB year). Accessed 2026-10-08.
• [18] U.S. Bureau of Labor Statistics. Employment Projections 2025 to 2035 (news release USDL-26-1422, August 27, 2026). 2025 base year, projected 2035. Accessed 2026-10-08.
• [19] U.S. Bureau of Labor Statistics. AI exposure categories and 2025 to 2035 employment projections (ai-exposure-categories.xlsx). 2025 employment, 2025 to 2035 projections. Accessed 2026-10-08.

Tier 2: trade bodies, market research, company data (5)

• [20] Associated General Contractors of America. AGC and Sage, The 2026 Construction Hiring and Business Outlook. Member survey, November 4 to December 15, 2025. Accessed 2026-10-08.
• [21] Royal Institution of Chartered Surveyors. RICS artificial intelligence in construction report 2025. Global survey, Q1 2025 (published September 12, 2025). Accessed 2026-10-08.
• [22] Precedence Research. Artificial Intelligence (AI) in Construction Market Size. 2025 estimate, 2035 forecast. Accessed 2026-10-08.
• [23] Global Market Insights. Artificial Intelligence in Construction Market Size and Share 2026 to 2035 (Report GMI6024, August 2026). 2025 estimate, 2035 forecast. Accessed 2026-10-08.
• [24] Mordor Intelligence. AI In Construction Market Size and Share Analysis, 2026 to 2031. 2025 and 2026 estimates, 2031 forecast. Accessed 2026-10-08.

Tier 3: journalism and expert commentary (1)

• [25] Construction Dive. AI nears 'tipping point' in construction as contractors pilot tech: survey (Construction Dive, December 10, 2025). Dodge Construction Network and CMiC survey, September and October 2025. Accessed 2026-10-08.