While the current housing shortage is a dire issue, it is merely symptomatic of a broader deficiency across American energy, manufacturing and infrastructure. However, this scarcity acts as a precursor to a new age of abundance, catalyzed by the emergence of artificial intelligence and its projected $3.5 trillion market value by 2033.
Of all concerns regarding construction in America today, the most frequently cited topic relates to the current housing shortage. And to be sure, the scarcity of affordable housing is a dire issue that demands an urgent remedy. But clearly there is an overarching concern that eclipses even the critical nature of this immediate deficiency in building.
In fact, what appears to be the general problem with homebuilding is only symptomatic of a greater series of overlapping and interrelated issues negatively affecting the construction industry, and the wellbeing of the nation.
Catalyst for a New Era
Somewhat akin to the way that residential construction inevitably translates into a plethora of related commercial building activities (only much more pervasively), a highly anticipated wave of overwhelming prosperity is toeing at the edge of a tidal surge. Ironically enough, the main thrust of this prescient wave lies with the undercurrent of prolonged scarcity.
Along with a shortage of adequate housing, the country currently faces deficiencies in energy, technology, manufacturing, transportation, infrastructure, and general innovation. In short, America requires, and demands, more of everything. Truly, there is a deepening hunger that is gnawing at the very core of our country.
But as the well-worn adage advises, necessity is the mother of invention, and the nation seems on the cusp of climbing out of a hole that it’s been digging for decades.
The good news speaks about how uniquely positioned these various economic sectors have become in affecting each other—like planets falling into alignment in a positive orbit. In fact, the foreshadowing that the current housing shortage portends, is likely the mere herald of the dawning of a new age of abundance.
The catalyst behind this latent birth of prosperity lies behind the advent of artificial intelligence (AI) and the promise (and demands!) it poses upon its inevitable emergence. The power emanating from this nascent emergence will undoubtedly spawn an unprecedented chain of events among several critically interrelated industries, as each industry affects its sibling sector.
Thus, the development of the intelligence product requires manufacturing (read, chips), which in turn requires data, which in turn requires data storage, which in turn requires construction and energy, which in turn requires infrastructure and transportation—all of which requires the innovation and creativity that explores, and continues to drive and develop (and invent) all the rest.
Recognizing that AI is the all-consuming, data-hungry monster that it portends to be, this analysis will initially focus its starting point on the raw material that the generation of the intelligence product requires. That primary scope is on manufacturing and construction. But even as a preface to the development of those initial industries, one must (however reluctantly) cite the statistics of the huge preliminary capital investment that AI is generating.
IT powerhouses such as Microsoft, NVDIA and QTS have pooled together with investment giants like Blackrock to help underwrite the anticipated growth of the AI market. Mind-numbing as they are, we cannot dismiss the early figures and their likely projections. And they are staggering. As of the latest available data, the current global AI market is worth almost $400 billion (U.S. market alone is worth upward of $47 billion).
What is even more stunning is the anticipated growth, which is expected to multiply nine times (at a compound annual growth rate of 31.5%) to land at a phenomenal $3.5 trillion by the year 2033. Of course, these numbers only reflect the current trajectory of the AI market value and are not necessarily predictive. Nevertheless, they are strongly indicative of an explosion of economic activity.
Reshoring the Silicon Foundation
That activity begins with manufacturing as the initial starting point of an ever-evolving chain of interrelated industrial sectors. And manufacturing, in relation to AI, equates with the production of chips. Of course, for some of us, the term still adheres to a spicy culinary product that is complemented by salsa. But a chip in the hyper-real sense refers to the tiny silicon-based microprocessor that is the building block of nearly all electronic functions.
As with many of the industries relating to our abundance model, U.S. chip manufacturing (that is, actual fabrication) has been in decline. While the U.S. semiconductor industry commands over 50% of global chips revenue (due to its superior strength of design and research and development), actual domestic fabrication has fallen from 37% in 1990 to a mere 10% currently.
This anomaly can be accounted for in large part by offshore fabrication and by incentivized foreign competitors. East Asian nations currently account for 75% of chip manufacturing worldwide, notably led by the People’s Republic of China.
Naturally this aggressive manufacturing activity has attracted uneasy recognition in the United States, which has, in part, given rise to a new wave of domestic incentives. National security is a recurring theme regarding chip manufacturing, and the federal government has not been reluctant to step in to remedy the recent shortfalls.
The Chips and Science Act of 2022 was passed to authorize $280 billion in funding to boost American semiconductor manufacturing and scientific research. The act, among other incentives, provides over $50 billion to direct domestic semiconductor development and manufacturing, with the express goal of reducing reliance on foreign suppliers.
Parallel federal programs designed to bolster chip manufacturing include the Department of Defense’s Secure Enclave Program, and a more recent federal move to “rescue” Intel (the nation’s top domestic chip producer) by publicly investing $8.9 billion for a 10% stake in the chip giant’s operations.
Not to be outdone by the government subsidized programs, private power players are jumping into the manufacturing game. Taiwan Semiconductor Manufacturing Company, which had already nearly completed a Phoenix-based facility, announced a $100 billion expansion at the same site. Similarly, Nvidia, which had previously kept most of its production off shore, has recently initiated plans to significantly increase its domestic footprint, earmarking supercomputer manufacturing expansions over the next four years in Texas and California to the tune of $500 billion.
This foregoing summation involves only the hardware aspect of the intelligence product. It only vaguely implies the circular relationship between the building blocks of all information technology, i.e., microprocessors (chips), and the generation of data. The hyperactive creation of data is a relatively new development, but a phenomenally explosive one, to be sure. In fact, the volume of data created, captured and consumed has exploded over the past five years, growing from 64.2 zettabytes (ZB) in 2020 to 180 ZB in 2025. Better said, more data has been created in the past three years than in all previous history. What’s more, global data generation is projected to more than double from 2025 levels by 2028.
The Infrastructure of Information
All of this data accumulation translates into a nearly insatiable need for data storage, and a corresponding need for the infrastructure to accommodate that storage. Data centers are the physical depositories for the storage, processing, and delivery of the immense amount of data generated by the intelligence product. Construction activity surrounding data center development has been nothing short of astonishing. There are 3,561 data centers operating in the United States today. That number is 17 times greater than it was a mere five years ago. But statistics tell only part of the story. The actual size and scope of this construction activity boggles the mind.
A recent Microsoft data center design typically encompassed a footprint roughly the size of three football fields. That spatial image translates into a couple hundred thousand square feet of drywall partition and a consistent magnitude of associated work per project. And while this current picture is optimistically indicative of remarkable levels of construction activity, the outlook for the near future promises to outpace the present, perhaps even exponentially so.
There is no question that the United States is currently experiencing a record-breaking data center construction boom, driven primarily by demands related to AI and associated cloud services. In 2025 year-to-date starts are projected to exceed $14 billion, with an average cost of $220 million per facility. Notable new projects include a 20 million square-foot data center in Buckeye, Arizona; a $2 billion campus in Sunbury, Ohio; a 137-acre Vantage Data Center campus for Storey County, Nevada; five newly added Oracle Stargate sites in Abilene, Texas; and a $600 million Google expansion in The Dalles, Oregon, among many others. In fact, no less than 78 data center facilities were under construction as of this writing.
But, as with any dynamic enterprise, there are bound to be serious obstacles, and the projected expansion of data centers is by no means an exception to the rule. Spatial restraints, for instance, are a primary challenge to unencumbered growth. Securing large plots of land is difficult and expensive. That demand for three football fields worth of real estate for each facility is a daunting proposition alone. But such constraints are not insurmountable. Better design for equipment racks that allow for more efficient cooling will help minimize space. Research on weight reduction can pave the way to accept stacking and the ability to “build up, not out.” And efforts to relocate tech centers to new regions with more abundant and less restricted land resources are currently underway.
But lack of affordable land is not the only obstacle looming over the further development of data centers. Not by a long shot.
Powering the Intelligence Revolution
The most foreboding barrier to unrestricted growth in this sector is one that provides a segue into another industry-wide deficit. The recurring theme of shortfalls in demand due to scarcity continues, this time with the generation and supply (or lack thereof) of power—specifically electrical power and how it interfaces with artificial intelligence.
Data centers alone could account for up to 12% of the nation’s total electricity load by 2028, up from 4.4% in 2023. The International Energy Agency states that global electricity demands from data centers could double by 2026. AI’s surging power demand is straining the already-feeble grid system in the United States.
The deterioration of America’s power grid can be characterized as alarming. Recent research indicates that 30% of transmission lines, and 45% of distribution equipment are near or beyond expected service life. Clearly the status quo has become unsustainable. Grid growth must match the demands of AI innovation, but the planned supply falls woefully short of the need.
Ironically, AI itself can become part of the solution to this shortfall using smart technology to manage the flow of electricity in more efficient ways and by balancing energy supply and demand to reduce transmission losses. Then too, existing transmission lines can be upgraded or rebuilt. But clearly, the optimal solution to the deficit is to build more power generating plants to increase supply and to retire/replace obsolete facilities.
But the creation of more plants requires better planning of precious energy resources, and that translates into changes in energy policy. A recent increase in production of fossil fuels, and a diminishing emphasis on renewable sources, has underscored a marked shift in energy policy. But even more particularly, energy policymakers are highlighting the importance of natural gas in the short term, and a preference for nuclear in the long term.
The justifications for preferring natural gas are several: better for the environment (burns cleaner than coal/oil, emitting less carbon dioxide), more stable (reliability for continuity of function with AI) and natural gas can be more cost efficient than other fossil fuels in the right context. The inclination toward nuclear is self-explanatory: it is the near-permanent solution to ongoing energy demand, and it eliminates greenhouse gases.
This chalks up energy generation as another critical industry that is currently plagued with deficiencies. Some efforts to shore up their weaknesses are ongoing, as noted, but have a lot of ground to make up. A related deficit that threatens the growth of AI lies with transportation, and in the age of AI, transportation equates to something more than planes, trains and automobiles.
Transportation means supply-chain management. Infrastructure, particularly energy infrastructure for the construction and operation of new data centers, depends on the effective functioning of vital supply chains. Consider that each new data center comprises a sovereign industrial project involving huge amounts of concrete, steel, copper and electrical equipment, all supported by multi-layered supply chains. Computation-hungry fabrication centers are drawing heavily on global manufacturing and logistics networks.
As with all the industries in our abundance model, scarcity becomes the common thread. In the case of supply-chain resilience, power is both the deficiency and the constraint. Project schedules depend heavily on when utilities can deliver the hardware for construction. But many significant projects have been put on hold due to lack of switchgear, transmission lines, substations, cooling systems and other related power-conversion hardware. And power-related capital costs are another serious obstacle to the free flow of supply.
Reliable generation of power requires billions of dollars, which translates into long lead times and delayed procurement, which in turn requires additional cost as the vicious circle goes on anon. And while initial construction costs may be relieved by private investors, the ongoing generation of power will ultimately fall upon the consumer in the form of utility rates.
Here it is painfully obvious that our model of interdependent industries has fallen under a rather dark theme of gloom and doom. The general notion of commerce to date seems to underscore a flock of core deficiencies. But, to steal a line from Gershwin, “it ain’t necessarily so.” And as bleak as the outlook may seem, the prospects for a rebound are uplifting—perhaps even inspiring. As has been strongly implied previously, scarcity is the harbinger of prosperity, just as necessity is the mother of invention. It should come as no surprise then, that the dawning of an age of prosperity should come on the heels of a period of scarcity and economic malaise.
Overcoming Bureaucratic Constraints
The cycle of scarcity has been underscored by decades of bureaucratic constraints. Blockers, not builders have been empowered to limit growth and innovation with untold volumes of restrictive measures and red tape that chokes the free flow of prosperity. Our original subtopic, the recent housing shortage, exemplifies the adverse effect of over-regulation. Restrictive zoning, i.e., setting a limit on how much housing can be put in a particular area, has created a mass exodus from significant population centers to those with more favorable, supply-side, economic climes.
Droves of people are “voting with their feet,” as restrictive zoning drives costs upward. California and New York are thrusting people out of their prospective states, while Texas and Florida are welcoming people with new cheap houses. And it’s not just the sunshine belt that has benefitted from zoning reform. Arizona, Montana, Idaho and Utah have all realized the advantages of deregulation. In fact, Utah, the most lightly regulated state in the nation, has created reform legislation in the form of the Environmental Permitting Modifications Act that has become the model for many of these other deregulated states.
But housing is not the only (nor most significant) industry that has been harmed by overly aggressive regulation, not by any stretch. As previously emphasized, the growth and innovation required for the emergence of AI needs cheap, reliable and abundant energy. However, landmark legislative measures, such as the Council on Environmental Quality, the Endangered Species Act, and the National Environmental Policy Act (NEPA) and other administrative bottlenecks have seriously stifled the growth of energy related projects.
In fact, research has surfaced to show that up to 32% of recent energy project cancellations have been related to excessive litigation during NEPA review. An ongoing spate of executive orders have been created to counter the negative impact of such over-regulation, and to streamline progress on energy-related advancements. In fact, a national energy emergency has been declared to cut through administrative red tape to expedite energy infrastructure projects, emphasizing a preference for natural gas and nuclear power.
Plainly, energy dominance is critical to AI, perhaps existentially so. Nuclear is the only path to reach the goal of affordable, plentiful, reliable, and clean energy. Nevertheless, the very mention of nuclear energy for many conjures images of Chernobyl and Three-Mile Island. But to meet the energy mandate that AI poses, Americans must move beyond aberrant catastrophes by rejecting taboos. Large, optimally located nuclear reactor sites must begin construction immediately due to long lead expectations, while small modular reactors (SMRs) and advanced nuclear reactors (ANRs) should be interspersed nationwide to be brought online, expediently.
Fortunately, there is a sunny side to the energy story (and one that’s not solar-related). TerraPower, a Bill Gates enterprise, has just broken ground on a first-ever, next gen, Natrium powered nuclear plant in Kemmerer, Wyoming. And the Triso-X energy corporation has begun construction in Oak Tennessee on a fuel fabrication facility for SMRs. And as for the interim, there are currently over a hundred natural gas plants in the early stages of development across the country.
Such anecdotal evidence indicates that a positive turnaround toward abundance is already occurring. AI mandates require nothing short of a major renaissance and that genie is already out of the bottle. And while scarcity will always be a driving factor toward problem-solving activity, we need only to look at what the human endeavor has achieved thus far. Clearly, abundant prosperity is imminent. CD
Vince Bailey is an estimator/project manager in the Phoenix area.
