Skip to content
AI Data Payout

Guides

Will AI replace architects?

AI is not on track to replace architects any time soon. Federal projections have architect jobs growing 4% by 2035, and in a 2026 test AI models read text on drawings well but counted doors and windows correctly only about half the time. Some drafting jobs are more exposed, with mechanical drafters projected to fall 6%.1,2,3

By Lead analyst (to be named) · reviewed by Reviewing editor (to be named) · published 1 Nov 2026 · updated 1 Nov 2026

What the job numbers say about architects

US architect jobs are projected to grow, not shrink. The Bureau of Labor Statistics counts 124,600 architect jobs in 2025 and expects 5,300 more by 2035, about 4% growth, with roughly 6,900 openings a year. Its handbook already says AI is part of why design work is getting more efficient, so that forecast is made with AI in view.1

Licensing also shapes the job. The same handbook says all states and DC require architects to be licensed, which ties the work to a person as well as to a drawing. A forecast is still a forecast, and nobody can see past the next few model releases, but the official number points up.1

What AI can do in architecture today

AI is already good at reading and writing around the drawing. In AECV-Bench, a January 2026 test of AI models on real floor plans and drawing questions, the best results on reading text and answering document questions reached 0.95 accuracy. In our reading that means real help with specs, notes and looking things up on a sheet.2

Use is spreading unevenly. In the US, an AIA study published in March 2025 found 6% of the profession regularly use AI for their jobs and 8% of firms had built it into practice. In the UK, RIBA found 59% of practices using AI in 2025, up from 41% a year earlier. Those surveys ask different questions in different countries, so read them as a range, not a trend line.4,5

What AI still can't do well

AI models still struggle to read a drawing as a drawing. The same test scored them at about 0.40 to 0.55 on counting doors and windows, and spatial reasoning was only moderate. The authors say today's models work as document assistants but lack robust drawing literacy, and they point to keeping people in the loop.2

Architects themselves are not expecting a takeover. In RIBA's 2025 survey, 18% of practices expected AI to lead to job losses, 35% worried it threatens the profession, and only 4% thought human creativity would stop being needed for building design. One test and two surveys can't settle the future, and models improve fast, so these numbers will date.5,2

Will AI replace CAD designers and drafters?

CAD designers and drafters are more exposed than architects, but the picture depends on the specialty. BLS projects drafters overall up 1% by 2035: architectural and civil drafters up 5% to about 109,500 jobs, electrical and electronics drafters down 3%, and mechanical drafters down 6%.3

The pressure BLS names is software that lets engineers and architects do work drafters used to do, which it pins on CAD and BIM rather than AI. AI could add to that for routine sheets, though that part is our reading and not a forecast. Drafters who check, coordinate and catch mistakes are doing the kind of work the January 2026 test shows AI is weakest at.3,2

Why firm archives have value

The gap in AI's drawing skills is also the reason firm archives are worth something. Fast Company reported that the data that could train an architecture model isn't widely available online, because drawings, 3D models and sketches sit on individual firms' servers and in filing cabinets. Our reading is that models stay weak at drawings partly because so few real ones are available to learn from.6

Buyers say the demand is real, though these are their own claims. Ventra says labs look for design files and CAD libraries specifically, and that one lab partner asked for 10,000 building designs. Polyshares names architecture and construction firms among the industries in demand. Licensing means handing over a copy, and whether a given drawing is the firm's to license is a question for the buyer you're working with.7,8,9

Pros and cons

Pros

  • Routine reading and writing around a project can move faster, and 84% of AIA respondents are optimistic AI can automate manual tasks to save time.4
  • Openings remain steady, at about 6,900 a year for architects and about 14,400 a year for drafters.1,10
  • Decades of drawings, which used to be dead storage, now have buyers.7,9

Cons

  • 90% of AIA respondents are concerned about AI risks such as inaccuracies, security and authenticity.11
  • Mechanical and electrical drafting are projected to shrink, even with no AI named as the cause.3
  • Models improve quickly, so a 2026 accuracy score is a snapshot.2

What it means for a business like yours

If you are an architect, the numbers say keep working and learn the tools. The work AI handles today is the reading and paperwork around design, and you hold the license and the judgment.1,2

If you are a CAD designer, your specialty matters more than the word drafter. Architectural and civil work is projected to grow, while mechanical is projected to shrink, so checking and coordination skills are the ones to build.3

If you run a firm, the same gap means your archive may be worth a conversation. Ventra told us CAD datasets don't need its 30-person minimum, and the quiz shows which partners take a firm your size.12

Should your business get in now?

Businesses are already licensing their records: micro1 has committed $1B over 12 months to buying company data, and at least 18 firms started advertising for it between September and mid-October 2026.13,14 Our view, not a promise: the money will not be around forever, because prices tend to come down as more businesses sell.13,15 Take the quiz to see which firms fit your business.

Plus a free Seller's Guide when you finish.

We feature companies that pay us a referral fee. Our scores and reviews don't change for it. How we earn.

FAQ

Questions, answered with sources

  • Nothing in the data points that way so far. BLS projects 4% job growth to 2035, and only 4% of practices in RIBA's survey think human creativity will stop being needed. Forecasts can change as models improve.1,5

  • CAD work is more exposed than architecture, by specialty. BLS projects architectural and civil drafters up 5% and mechanical drafters down 6% by 2035, and it ties that pressure to CAD and BIM software.3

  • CAD software looks more likely to absorb AI than to disappear, and that is our reading. The January 2026 test found AI models weak at reading symbols on drawings, at about 0.40 to 0.55 on counting doors and windows.2

  • Figures vary by survey. AIA found 6% of the US profession regularly use AI, in a study published in March 2025, and RIBA found 59% of UK practices using it in 2025.4,5

  • Reading a drawing reliably is the clearest gap. Models in AECV-Bench handled text well but scored about 0.40 to 0.55 on counting doors and windows, and the authors say they lack robust drawing literacy.2

  • AI companies want drawings because good examples are hard to find. Fast Company reported they sit on firms' own servers instead of the open web, and Ventra says labs look for CAD libraries specifically.6,7

  • Every business is different, so this one is best talked through with the buyer you're working with. Their team goes over it with you before anything is signed.

  • Every business is different, so this one is best talked through with the buyer you're working with. Their team goes over it with you before anything is signed.

Before you sign

Every business is different, so these are best talked through with the buyer you're working with. Their team goes over them with you before anything is signed.

Questions for this kind of business

  • Which project files does our firm hold the rights to license, and which are the client's?
  • Do our design software terms say anything about training AI on our files?
  • Who owns this data: the company, an estate, or someone else?
  • Do customer, supplier or platform contracts, NDAs or API terms restrict this?

Questions for any buyer

  • Is the license exclusive, and for how long?
  • Can you sublicense the data, and to whom?
  • Which AI labs will receive it, and will you tell us?
  • How long do you keep it, and how is it deleted?
  • What is your liability cap, and who indemnifies whom?
  • Will the contract ban re-identification of people in the data?
  • Who removes personal data, by what method, and do we approve it first?

How this page was made: drafted with AI help from the sources below, each checked; a buyer's claims about itself are its own. Updated November 2026. How we rank.

Sources (15)

  1. 1.US Bureau of Labor Statistics, Occupational Outlook Handbook. Accessed 1 Nov 2026.
  2. 2.Kondratenko et al., AECV-Bench, arXiv, published 8 Jan 2026. Accessed 1 Nov 2026.
  3. 3.US Bureau of Labor Statistics, Occupational Outlook Handbook. Accessed 1 Nov 2026.
  4. 4.The American Institute of Architects, published 11 Mar 2025. Accessed 19 Oct 2026.
  5. 5.Royal Institute of British Architects, published 27 Jun 2025. Accessed 1 Nov 2026.
  6. 6.Fast Company (Nate Berg), published 24 Aug 2026. Accessed 19 Oct 2026.
  7. 7.Ventra company data page. Accessed 15 Oct 2026. Company's own claim.
  8. 8.Ventra founder, call with AI Data Payout (2026-10-19), published 19 Oct 2026. Accessed 19 Oct 2026. Company's own claim.
  9. 9.Polyshares broker portal, market criteria update (2026-10-22), published 22 Oct 2026. Accessed 22 Oct 2026. Company's own claim.
  10. 10.US Bureau of Labor Statistics, Occupational Outlook Handbook. Accessed 1 Nov 2026.
  11. 11.American Institute of Architects, published 11 Mar 2025. Accessed 1 Nov 2026.
  12. 12.Kyle at Ventra, email to AI Data Payout (2026-11-01), published 1 Nov 2026. Accessed 1 Nov 2026. Company's own claim.
  13. 13.The Next Web, published 9 Oct 2026. Accessed 14 Oct 2026.
  14. 14.Meta Ad Library (our review). Accessed 14 Oct 2026.
  15. 15.TechCrunch, published 16 Sep 2025. Accessed 14 Oct 2026.