Recently, Global Location Strategies was working with an industrial user evaluating a site in South Carolina. The initial land proposal came in at $90,000 an acre, high by historical standards. A few weeks later, before we had taken a single next step, the price had risen to $180,000 an acre. Same site. Same dirt. The only thing that had changed was the fact that a data center developer had expressed interest.
We did not lose that site to a competing manufacturer. We did not lose it to a better-prepared company or a more competitive region. We lost it to a different industry entirely: one with a fundamentally different cost structure, a different relationship with power and a different tolerance for what land and infrastructure are worth.
That is the conversation the automotive manufacturing industry is not having. The industry is focused on EV demand uncertainty, tariff exposure, OEM capital discipline and the execution constraints that have always made vehicle assembly plants, battery facilities and supplier investments difficult. Those are real issues. But the constraint that will determine which automotive investments get built in the next five years and which ones quietly stall in site selection is power. The competition for power has changed in ways that most automotive capital allocation processes have yet to catch up with.
The AI data center buildout is consuming power capacity, locking up land and absorbing construction resources at a scale and speed that the U.S. infrastructure system was not designed to absorb. Nearly $930 billion in data center capital expenditure is being compressed into roughly six years. For context: the Interstate Highway System cost $620 billion (adjusted for inflation) over 37 years. We have never moved this much infrastructure investment this quickly. Automotive manufacturers need the same power, the same sites and the same construction labor. They are now competing for all of it in a market in which they are not accustomed to competing.
The New Math of Power Cost
Here is the asymmetry that makes this competition structurally unfair, and why it matters for every automotive location decision being made right now.
A one-cent-per-kilowatt-hour increase in electricity cost for an automotive OEM producing 200,000 vehicles annually can add roughly $2 million in operating expense, or approximately $10 per vehicle. While that may sound modest, automotive manufacturing operates on thin margins and every additional cent compounds quickly.
A data center consuming a similar amount of electricity often distributes those costs across millions of users, transactions or AI workloads. The economics are fundamentally different. Automotive manufacturers and hyperscalers may compete for the same infrastructure, but they are not equally sensitive to changes in energy cost. Utilities are in the business of managing risk and ensuring returns. They are making decisions about who gets served first, who gets the more favorable interconnection timeline and whose load growth they plan around.
The answer, in a growing number of markets, is not the manufacturer, not because utilities don’t want to serve manufacturers, but because hyperscalers can commit to larger loads more quickly. Manufacturing site selection is typically not a race to be first. It is competition to see who can be most competitive over the long term.
This is not theoretical. In region after region, industrial customers may not just fail to show up. They may fail to expand. Energy prices could push them into cost-cutting mode, and labor and new capital investment are the first things cut. That is unrealized investment: jobs and capital that communities and companies never see because the economics quietly stopped working before anyone made a formal announcement.
When the Margin for Error Disappears
The power competition matters more today than it would have a decade ago because of what has happened to the scale and irreversibility of automotive capital commitments. Understanding that shift is what gives the site availability problem its full weight.
The post-2020 automotive investment environment looks, on the surface, like a story of decline. Project counts are lower than they were in the mid-2010s. That comparison is misleading. The capital data tells a different story. According to FDI Markets, projects exceeding $1 billion accounted for 18 percent of total automotive capital expenditure in 2015. By 2025, that share had reached 43 percent, after peaking at 44 percent in 2022. Average capital expenditure per project has increased materially. Employment per project has grown more slowly, reflecting capital deepening into increased automation rather than proportional labor growth.
The practical consequence is straightforward: individual automotive investments now carry far more financial exposure than they did a decade ago. Delays in commissioning and extended ramp-up timelines no longer result in manageable slippage. They translate into prolonged periods of underutilized assets and deferred returns that are difficult to absorb on a corporate balance sheet. Capital risk rarely materializes at announcement. It surfaces in the interval between commitment and sustained production, when constraints in power delivery, workforce ramp or permitting alignment prove more complex than initially modeled.
This is the environment in which a site lost to a data center speculator, or a substation with no available capacity, becomes a balance sheet problem rather than a recoverable inconvenience. The margin for error has compressed at exactly the moment when the competition for the infrastructure automotive investments require has intensified.
Energy Disruptions Added to an Stressed System
Recent disruptions across global energy markets illustrate a broader point: manufacturers no longer operate in a world where energy is simply a domestic utility issue. While the United States remains relatively insulated compared to many regions, electricity prices and reliability are increasingly shaped by interconnected factors including fuel markets, liquified natural gas (LNG) exports, transmission constraints and geopolitical events. Energy markets move quickly. The structural implication will not.
When a major energy corridor is disrupted, it does not just move fuel prices. It creates uncertainty across shipping costs, manufacturing input costs, production timelines and long-horizon business planning. For manufacturers operating 50-year facilities under thin margins, that kind of uncertainty is not something that can be hedged away. It has to be underwritten by the location decision itself. That means the energy cost structure of a site, and the stability of that structure over time, have become more heavily weighted in location decisions and increasingly determine which sites remain competitive over the life of an investment.
The automotive industry was already experiencing this shift before recent energy disruptions emerged. The issue is not a single fuel source or a single geopolitical event. It is that the margin for error across the energy system has narrowed. Power demand is growing rapidly, generation and transmission additions take years to develop and new large-scale loads are being added faster than infrastructure can adapt in many markets.
Manufacturers have historically operated in a system with enough flexibility that power availability and cost assumptions were relatively stable over the life of an investment. That flexibility is becoming harder to assume. Recent disruptions did not create the constraint. They exposed and amplified one that was already developing.
For automotive executives evaluating North American manufacturing locations, the question is not whether energy markets will be volatile. They will be. The question is whether the location decision was built to absorb that volatility, or whether energy cost assumptions that were made at the announcement will hold through the payback period.
The Site You Think You Have May Not Be Available
The competition for power is not the only challenge. It is a site availability challenge, and it is moving more quickly than most automotive site selection timelines can track.
Data center development has changed the land market in ways that do not show up in traditional site databases. A site that was available, appropriately priced and power-credible 18 months ago may now be under option to a data center developer, repriced to reflect that interest, or simply no longer viable because the substation capacity it depended on has been allocated to a different load. The South Carolina example reflects a market where data center capital, which moves faster and tolerates higher land costs, has structurally changed the acquisition environment for industrial users.
The concentration dynamic amplifies this. Data center development clusters around power infrastructure, fiber networks and cooling resources. That clustering creates pockets of acute scarcity in markets that were previously considered reliable options for manufacturing investment. A region that was on an automotive shortlist three years ago may now have a fundamentally different power availability picture because 500 megawatts of new baseload demand showed up before the project did.
There is a secondary effect that runs in the opposite direction, and it is worth naming. In some markets, data center infrastructure investment has created conditions that benefit manufacturing. The New Albany Business Park outside Columbus is the clearest example: early data center development funded transmission upgrades and skilled trades capacity that later supported Intel, Amgen and continued industrial investment. The infrastructure the data centers paid to build became the infrastructure manufacturing used.
The lesson is not that data centers are the enemy. The lesson is that the relationship between data center development and manufacturing viability in any given market is not predictable from the outside. It requires ground-level knowledge of what infrastructure has been committed, what capacity remains and what the utility’s planning horizon actually looks like for industrial load. That intelligence is not available in a database. It requires direct utility relationships and real-time market data.
What Has to Change in How Automotive Location Decisions Are Made
The standard automotive site selection process was built for a different competitive environment. It screens for labor, logistics, infrastructure and incentives, in roughly that order, with infrastructure often treated as a confirmable detail rather than a primary filter. That sequence no longer reflects the actual constraint structure of the market.
Power has always been part of the fatal flaw screen. What has changed is that a growing number of sites are no longer passing that screen or cannot provide enough certainty that they will. That reality moves power to the front of the evaluation process. The questions that now need to be answered at the start of a site search, not during due diligence, are: What is the utility’s current interconnection queue, and how long is it? What load has been committed to the substation serving this site? Can the utility contractually validate power delivery against the commissioning schedule? What is the long-run cost trajectory for industrial customers in this territory, given the new load that has been added?
Sites that cannot answer those questions credibly should be removed from consideration before the comparative analysis begins. The organizational cost of removing a site that has built internal momentum during due diligence is high. The financial cost of committing capital to a location that cannot deliver power on schedule is higher.
Land acquisition timelines need to compress. The traditional site selection process moves at a pace that made sense when industrial manufacturers were the primary competition for large, infrastructure-ready sites. That pace is no longer safe. In markets where data center interest is active, the gap between site identification and site control has become a risk window. Sites go under option, get repriced or lose their power availability while corporate evaluation processes are still running their normal course.
The geographic assumption set needs to be rebuilt. The map has changed, and not in a single direction. Some historically attractive regions have become more constrained as power capacity has been absorbed, while other markets have become more competitive because infrastructure investment created new capacity. Evaluation processes working from prior assumptions about which markets are viable are working from an outdated map.
The New Constraint Structure
The automotive manufacturing investment environment has been reshaped by two forces operating simultaneously: the capital concentration that has made individual investments larger, less reversible and more execution-sensitive than at any point in the past decade, and the AI infrastructure buildout that has fundamentally changed the competition for power, land and the construction resources those investments require.
Neither of those forces is temporary. The capital concentration reflects structural changes in how automotive platforms are designed and funded. The data center buildout reflects demand for AI infrastructure that is unlikely to reverse. Both will continue to shape the environment in which automotive location decisions are made.
What is optional is whether automotive executives and their site selection advisors update their processes to reflect that reality. Companies that move power to the front of the evaluation, compress site control timelines and rebuild geographic assumptions using current intelligence rather than outdated assumptions will preserve flexibility and choice. Companies that do not may discover the constraint during due diligence, when the cost of adjustment is high and the calendar pressure is real.
Automotive site selection used to be about identifying the best location among many viable options. Increasingly, it is becoming a race to secure one of the few locations capable of supporting long-term industrial growth.
The site you think you have may already be gone. The question is whether you find that out before you’ve committed. T&ID
About the Author
