Built Intelligence
Value ManagementRisk AllocationBIMGovernanceDigital TwinsAI & Construction 4.0Why It MattersTechnologiesGlossary

Why it matters

The National Picture

The argument of this practice rests on public numbers. This page assembles them: what the United States is building, what the known losses are, who is available to build it, and which rules already exist. Notes at the foot of the page give the sources; the six chapters treat each subject in depth.

1. The buildout

The most explicit federal construction priority of this moment is the infrastructure of artificial intelligence. Executive Order 14318, Accelerating Federal Permitting of Data Center Infrastructure (July 23, 2025), states it directly: “It will be a priority of my Administration to facilitate the rapid and efficient buildout of this infrastructure …” and defines a “Qualifying Project” as a data center requiring more than 100 megawatts of new electric load or $500 million of capital investment.1 The same day, America’s AI Action Plan devoted an entire pillar to “Build American AI Infrastructure,” including the workforce to build it.2 The Department of Energy has opened federal sites for AI data centers and their power; among the first selected is the Savannah River Site, in the Southeast.3 The market moved with the policy: private data-center construction reached an annual rate of roughly $50.7 billion by April 2026, passing office construction for the first time on record.4

Behind that front sits the earlier wave: $1.85 trillion authorized under the Infrastructure Investment and Jobs Act, the CHIPS and Science Act and the Inflation Reduction Act, still delivering semiconductor plants, grid work and federal modernization. A portion of the IRA’s clean-energy credits was ended early by legislation in July 2025, which is why this page leans on what is current rather than on a single headline number.5 All of it lands on an inventory in poor condition: the American Society of Civil Engineers grades overall U.S. infrastructure at C in its 2025 Report Card.6 Building this much, this fast, on this base, is above all a test of delivery structures: of how projects allocate risk, hold schedules and keep their information straight when everything is under pressure at once. That is Chapter 02 of this practice, and the power decision behind every one of those campuses is Chapter 01.

2. The demand shock: data centers

Lawrence Berkeley National Laboratory, reporting for the Department of Energy, measured U.S. data centers at “176 TWh by 2023, representing 4.4% of total U.S. electricity consumption,” up from about 76 TWh and 1.9 percent in 2018.7 Its scenarios for 2028 span “the low and high end of roughly 325 and 580 TWh”: “6.7% to 12.0% of total U.S. electricity consumption forecasted for 2028,” a load the report translates into “a total power demand for data centers between 74 and 132 GW.”7 The Electric Power Research Institute points the same direction on a longer horizon, with data centers reaching “4.6% to 9.1% of U.S. electricity generation annually by 2030,” and notes the concentration: “fifteen states account for 80% of the national data center load, with data centers estimated to comprise a quarter of Virginia’s electric load in 2023.”8 Every one of those facilities is a capital project whose power-supply decision is made once, early, and lived with for decades. What a disciplined version of that decision looks like is the worked model in Chapter 01.

3. The productivity gap

Construction is one of the largest sectors of the world economy, with about $10 trillion a year in spending, and its productivity record is the industry’s best-documented failure. The McKinsey Global Institute’s Reinventing Construction put the global productivity opportunity at $1.6 trillion a year, enough to meet “about half of the world’s annual infrastructure needs,” and found that the U.S. construction sector “accounts for one-third of the opportunity to boost global productivity identified in this research.”9 The comparison inside the U.S. economy is stark: since 1945, productivity in manufacturing, retail and agriculture has grown by as much as 1,500 percent, while construction’s labor productivity “is lower today than it was in 1968.”9 The baseline on individual projects is documented separately: large projects across asset classes “typically take 20 percent longer to finish than scheduled and are up to 80 percent over budget.”10 MGI’s answer is a list of seven actions that together could raise sector productivity by 50 to 60 percent, and one of them is squarely this practice’s subject: “rewire the contractual framework to reshape industry dynamics.”9 How contracts do that rewiring is Chapter 02.

4. The information loss

The National Institute of Standards and Technology quantified what bad information costs the U.S. capital facilities industry: $15.8 billion a year in 2002 dollars, from inadequate interoperability alone.11 The distribution is the sharper finding: roughly two-thirds of the burden, about $10.6 billion, falls on owners and operators, and most of it is paid during operations and maintenance, year after year, by whoever inherits the building.11 The failure persists even on well-intentioned projects: the peer-reviewed record documents a $113 million research facility with a formal BIM execution plan and required handover data whose automated exchange into the maintenance system failed anyway.12 Preventing exactly that is Chapter 04, and keeping the information alive afterwards is Chapter 05.

5. The people shortage

The industry building all of this is short of people. Associated Builders and Contractors’ workforce model estimates the industry must attract 349,000 net new workers in 2026 just to meet demand, rising to 456,000 in 2027 as spending growth resumes, on top of normal hiring.13 The shortage compounds every other number on this page: a workforce that scarce cannot afford rework, disputes, idle crews or lost information. Every hour that better-governed delivery recovers is an hour of national capacity returned, which is why the workforce case for disciplined delivery is not separate from the productivity case: it is the same case, counted in people instead of dollars.

6. The rules already on the books

None of this waits for new law; the rules largely exist. The General Services Administration established its national building-information program in 2003 and made a spatial program model “the minimum requirement” for major projects from fiscal year 2007; its current facilities standard requires model deliverables “throughout the project period, at multiple milestone points,” with an open IFC submission alongside each native one.14 The U.S. Army Corps of Engineers is categorical: advanced modeling is “required as described herein” on military projects over 5,000 gross square feet and $3 million and on civil works over $3 million, and its 2026 engineer manual carries the requirement through construction records.15 Value engineering is statute: every executive agency “shall establish and maintain cost-effective procedures and processes for analyzing the functions” of its programs and facilities, and OMB Circular A-131 requires VE for projects of $5 million and up.16 For artificial intelligence, NIST’s AI Risk Management Framework has structured the risk questions since 2023,17 and ISO/IEC 42001 gives organisations an auditable AI management standard.18 The mandates are treated with full citations in Chapter 04, Chapter 01 and Chapter 06.

7. What follows

Put the pieces together. An unprecedented volume of construction is funded and underway. The industry delivering it has a documented productivity problem, a documented information problem, and a documented shortage of people. And the standards that address all three already exist, in public, many of them mandatory on public work, but on most projects they have not been turned into decisions anyone is obligated to make. That last step, from standard to decision, is where this practice works, in open access, for the small and mid-sized firms that carry most of the industry’s output. The story runs in six chapters, starting with what the money must buy.

Sources

  1. Executive Order 14318, Accelerating Federal Permitting of Data Center Infrastructure (July 23, 2025).
  2. Winning the Race: America’s AI Action Plan (July 23, 2025), Pillar II, “Build American AI Infrastructure.”
  3. U.S. Department of Energy, site selections for AI data center and energy projects on federal lands (2025), including the Savannah River Site.
  4. U.S. Census Bureau, C30 Construction Spending, data center category, seasonally adjusted annual rate, April 2026.
  5. Infrastructure Investment and Jobs Act, CHIPS and Science Act, and Inflation Reduction Act; authorized amounts per Congressional Budget Office figures. Several IRA clean-energy credits were terminated early by the reconciliation act signed July 4, 2025.
  6. American Society of Civil Engineers, 2025 Report Card for America’s Infrastructure.
  7. Shehabi, A., et al., 2024 United States Data Center Energy Usage Report, Lawrence Berkeley National Laboratory, LBNL-2001637 (December 2024).
  8. EPRI, Powering Intelligence: Analyzing Artificial Intelligence and Data Center Energy Consumption (2024).
  9. McKinsey Global Institute, Reinventing Construction: A Route to Higher Productivity (2017), Executive Summary.
  10. Agarwal, R., Chandrasekaran, S., & Sridhar, M., Imagining construction’s digital future, McKinsey & Company (2016).
  11. NIST GCR 04-867, Cost Analysis of Inadequate Interoperability in the U.S. Capital Facilities Industry (2004).
  12. Pishdad-Bozorgi, P., Gao, X., Eastman, C., & Self, A. P., Planning and developing facility management-enabled building information model (FM-enabled BIM), Automation in Construction 87, 22–38 (2018).
  13. Associated Builders and Contractors, construction workforce model (January 2026).
  14. GSA, National 3D-4D-BIM Program (est. 2003); GSA BIM Guide Series 01 (2007); P100 Facilities Standards for the Public Buildings Service (2024).
  15. USACE, ECB 2018-7 Rev. 3, Advanced Modeling Requirements on USACE Projects (2025); EM 1110-1-4017, Advanced Modeling for Digital Project Delivery (2026).
  16. 41 U.S.C. §1711; OMB Circular A-131, Value Engineering (2013).
  17. NIST AI 100-1, Artificial Intelligence Risk Management Framework (AI RMF 1.0) (2023).
  18. ISO/IEC 42001:2023, Information technology — Artificial intelligence — Management system.

Terms used on this page

Short definitions for the acronyms and terms of art above. The complete vocabulary is in the A–Z glossary.