These nontech stocks can ride the AI build-out for years
Source Entity
Philip van Doorn
The AI infrastructure boom is driving massive investment into data centers, but labor shortages and 'phantom' projects are complicating growth. While 'HALO' companies stand to benefit, grid operators face significant challenges in accurately forecasting power demands.
The Double-Edged Sword of the AI Infrastructure Boom
The rapid proliferation of artificial intelligence has triggered an unprecedented surge in capital expenditure, with hyperscalers projected to deploy roughly $800 billion by 2026 and exceeding $1 trillion by 2027. This massive investment is not limited to software; it is fundamentally transforming the physical landscape of the United States as Big Tech firms race to establish data centers. This expansion creates unique opportunities for what analysts term 'HALO' companies—those characterized by hard assets and low obsolescence—which are essential to the physical build-out of these facilities.
The Labor Bottleneck: A Crisis of Skill
Despite the staggering financial commitments, the physical construction of these state-of-the-art facilities faces a critical obstacle: a chronic shortage of skilled human labor. While the U.S. mechanical, electrical, and plumbing (MEP) workforce is often cited at 1.8 million, Bernstein analyst Chad Dillard warns that this figure is misleading. A large portion of this workforce is either geographically inaccessible, unavailable, or lacks the specialized certifications required for high-tech data center construction. With only 790,000 workers currently in the non-residential sector, the industry is struggling to find the welders and electricians necessary to meet the aggressive timelines set by tech giants.
The Rise of 'Phantom' Projects
Compounding the construction challenges is the emergence of 'phantom' data center projects. These are initiatives that are either announced without the intent to build or are proposed at scales that ignore the physical constraints of the electrical grid. This phenomenon has become a primary disruptive force, making it nearly impossible for grid operators to forecast actual power demand. When utilities cannot distinguish between speculative announcements and concrete projects, the resulting uncertainty leads to inefficient resource allocation and grid instability.
Economic Implications and Grid Strain
The strain on the electrical grid is already manifesting in the consumer sector. As anticipation for high data center loads drives massive infrastructure investment, household electricity prices have risen by 10.1% over the past two years, outpacing broader inflation metrics. The power requirements are immense; companies like Nebius and CoreWeave are rapidly scaling their contracted and active power needs, with CoreWeave expanding its active power by nearly 500 megawatts in a single quarter. This creates a competitive environment where AI ambitions directly clash with the energy needs of the general population.
Future Trends and Strategic Outlook
Looking ahead, the success of the AI build-out will likely hinge on the industry's ability to reconcile its infrastructure goals with reality. We can expect increased regulatory scrutiny regarding energy consumption and a greater emphasis on workforce development to bridge the MEP labor gap. Furthermore, as grid operators gain more transparency into project pipelines, the market may see a cooling of speculative 'phantom' announcements in favor of more sustainable, data-driven expansion. Investors should focus on companies that provide the foundational hard assets required for these facilities, as they are best positioned to navigate the volatility caused by labor shortages and energy grid constraints.
Multiple Citing Sources
Verification Required?