For 25 years, US manufacturers had one priority: build where it is cheapest, which often meant abroad. That logic is breaking. According to our analysis, US companies are planning $6 trillion in domestic investments between 2025 and 2029. That capital—concentrated primarily in data centers and AI compute, along with the electrical grid—represents a roughly threefold increase compared with investments over the past five years. It’s a once-in-a-generation opportunity.
Yet even though capital is easier to access, capacity is not. Manufacturers face critical bottlenecks in talent, supply chains, and assets. Each will take years to address, meaning that companies must act quickly or risk missing the window. Asset productivity tells a similar story: gross profit per dollar of net property, plant, and equipment (PP&E) rose from the mid-1990s to roughly 2015 as digital tools modernized aging plants, then declined over the past decade as those gains matured, leaving room for a new wave of productivity gains today.
This is the challenge of our age, and the next industrial chapter for the US. It’s similar in scope to major economic transitions like the construction of the railroads, the electrical grid, the post-WWII manufacturing boom, and the internet. Companies that want to win in this environment must understand the driving factors of US reindustrialization, address the key constraints holding them back, and revisit their labor strategy to secure the workforce they need now and in the years ahead.
Ride the Supercycle of Capital Investment
The nearly $6 trillion to be injected into the US economy between 2025 and 2029 represents about one-fourth of all US investment over that period, and it will be concentrated in a few key sectors. Just two—data centers (including AI compute) and electrification (including power generation, transmission, and distribution infrastructure)—will make up about 85% of that amount. That’s roughly a threefold increase compared with investments over the past five years. (See Exhibit 1.)
Almost every industrial firm will be impacted. Some have direct exposure to these sectors, and they will succeed through speed and delivery certainty. But a far bigger pool of companies has indirect, downstream exposure through second-order demand. AI compute, data centers, and grid projects will all require components (such as steel, rare-earth magnets, and critical minerals); cooling systems; and power storage, transmission, and distribution solutions. And companies will be able to use AI and electrification to reinvent their business models and create value for customers in new ways.
This represents a potential flood of new business, but companies will need to adjust their operations accordingly. In particular, incumbents will have to do things they haven’t done in the past, moving faster than many of them are used to.
Map Your Exposure. Winners will identify where they have credible exposure to data centers, AI compute, electrification, and grid buildout. That entails looking far beyond legacy operations and determining where the company could make an intelligent pivot to capture adjacent demand. Companies may need stronger simulation and modeling capabilities to assess upstream impact before they have complete clarity.
Revamp Your Strategy. It’s not enough to rely on legacy brand strength or historically acceptable service levels. Instead, companies need to quickly pivot to tap into new demand. They must adapt their offerings and operating models to capture emerging opportunities; for example, developing new products for adjacent applications, adding production lines or shifts, securing critical inputs, and building the commercial capabilities to reach a different set of customers. In one such move, Ford is repurposing its EV battery supply chain to sell battery storage systems to utilities, data centers, and large industrial and commercial customers.
Accelerate Product Development. As AI compresses design, simulation, and testing cycles, speed is becoming a greater source of competitive advantage than ever. AI-native new entrants can move from concept to market-ready product faster than large incumbents. To compete for new business, companies need to embed AI into product design and engineering and focus on speed as well as quality.
Address Bottlenecks in Talent, Supply Chains, and Infrastructure
The wave of capital is coming when industrial manufacturers are less equipped to put it to use than in the past, due to structural constraints in three main areas: labor, supply chains, and infrastructure. Each of these bottlenecks can take years to clear, which means the companies that solve these issues fastest—both for themselves and their customers—can gain a durable advantage.
Labor: A Shortage of 2 Million Skilled Tradespeople. People may prove to be the toughest bottleneck in the buildout. US manufacturers face a talent shortfall of approximately 2 million skilled-trade jobs by 2030, putting an estimated $1 trillion of annual economic activity at risk. By 2035, the gap will remain acute across three trades: mechanical and plumbing, electrical, and instrumentation and control workers.
A complicating factor is that the US economy has been steadily reoriented toward knowledge work, including professional and business services, finance, insurance, and information and telecommunication services. Since 1975, this area has more than doubled its share of GDP, while manufacturing’s share has roughly halved. (See Exhibit 2.)
Over the long term, physical AI and robotics will likely ease labor constraints. But manufacturers can’t wait until that happens. In the near term, they can generate faster results by investing in proprietary training, upskilling, and automation to improve the productivity of their current workforce. For example, Toyota’s Advanced Manufacturing Technician program (created in 2005) trains job-ready, multiskilled maintenance technicians in two years; 95% are hired full-time by their sponsoring company, and 83% finish on time.
Supply Chains: Long Lead Times for Key Components. The buildout cannot run on today’s supply chain. A wide set of critical inputs and components, mostly at the intersection of AI compute and power, are concentrated in constrained supply chains, and many face multiyear backlogs.
For example, demand for electric-grade silicon steel (a core material in transformers) is up threefold since 2015. Similarly, the liquid-cooling systems used in data centers require highly engineered components such as cold plates, cooling distribution units, and manifolds, most of which are produced at a small number of tier-2 suppliers.
To reduce supply chain constraints, manufacturers need a multipronged approach. In some cases, they can qualify and support additional suppliers in the US. Standardizing designs can help as well, ultimately reducing product complexity, attracting new players, and creating redundant sourcing options.
Other materials, such as some critical minerals, cannot be sourced domestically. Instead, manufacturers may need to source from allies, along with recycling and recovering inputs from existing mineral streams. And some companies are moving upstream in the value chain, through new facilities that can process or refine inputs. For example, Tesla built a $1 billion lithium refinery in Texas to bring battery inputs in-house. The facility can process approximately 50,000 tons per year of battery-grade lithium hydroxide.
Infrastructure: Aging and Underutilized Factories. Even if manufacturers could quickly solve the talent and supply chain constraints, they still face a third bottleneck: the existing industrial base is aging and underutilized, and scaling back to full production rates may be too costly. Consider:
- Industrial production utilization is currently about 75% in the US. That’s up from the 2009 low of 68% but still below the pre-2005 average of 81%.
- Assets are less productive. Gross profit per dollar of net PP&E rose through 2010 as companies used digital tools to modernize largely analog plants, but it has since fallen from $0.95 to $0.74, even as gross margins remained broadly stable. The divergence points to declining asset productivity and highlights the potential for a new wave of productivity gains enabled by digital technology and AI. (See Exhibit 3.)
- Building newer and bigger factories may not make economic sense given the rapid growth of construction costs. Since 2009, the Producer Price Index for nonresidential construction has increased by about 80%, compared with about 50% for the broader Consumer Price Index. Permitting timelines, litigation risk, regulatory uncertainty, and constrained labor and equipment availability all raise the hurdle rate on new capacity.
Instead of building new greenfield plants, a faster and more cost-effective approach is to modernize legacy plants. BCG’s recent Factory of the Future analysis found that technologies such as predictive maintenance, digital process control, and IOT can lift manufacturing productivity by as much as 60%. In some cases, upgrading a plant in a high-cost country such as the US can make it competitive with upgraded facilities in lower-cost countries.
In some cases, upgrading a plant in a high-cost country such as the US can make it competitive with upgraded facilities in lower-cost countries.
For example, Schneider Electric retrofitted a 60-year-old plant in Lexington, Kentucky. The company invested in industrial IoT connectivity, predictive analytics, and digital management tools, reducing energy use by 26% and equipment downtime by 20%.
Localize Value Chains to Prioritize Quality and Speed
The push to outsource manufacturing to low-cost countries, primarily in Asia, is reversing. Physical constraints, geopolitical tensions, growing supply chain risks, and other factors have pulled the center of gravity back to the US. (See Exhibit 4.) Evolving tariff policies are another factor. Section 232 and related trade actions have made landed-cost calculations more uncertain, and the next phase of the United States-Mexico-Canada Agreement (USMCA) could reshape which products qualify for preferential treatment.
As a result, US manufacturers want to produce goods closer to domestic customers. In a 2022 survey of 183 US manufacturing executives, 73% prioritized resilience and responsiveness, up from 32% in 2019. In a 2025 reshoring survey, 40% said they would pay a 10% to 20% premium to cut delivery times by five weeks. The shift is most visible in high-value and critical components, where longer lead times, quality risk, and uncertain supply undercut the value of cost savings.
In semiconductors, for example, more than 90% of leading-edge chips are produced in Taiwan. This leaves the AI compute sector on which the entire supercycle depends vulnerable to potential disruption arising from tensions over the Taiwan Strait. In response, Taiwan-based chip manufacturer TSMC is investing roughly $165 billion in the US, anchoring advanced chip production and packaging in Arizona.
Similarly, Mainland China accounts for roughly 90% of rare-earth magnet production, a critical input for data centers, power equipment, defense systems, EVs, and industrial machinery. MP Materials, a rare-earth materials company based in the US, is building a magnet supply chain outside Mainland China, underpinned by a $500 million commitment from Apple.
For other industrial manufacturers, it’s critical to revisit the footprint strategy through a lens of speed, quality, cost, and risk. Evolving tariff policies need to be in the equation as well. Leaders need to pressure-test whether their North American footprint stays economically viable under USMCA 2.0 before the terms land, identify nearshoring opportunities, and continue investing in automation, advanced analytics, and digital operations, to continue closing the cost gap with other regions.
Streamline the Portfolio and Focus on Winning Businesses
The fourth and final theme is strategic focus. Capital is abundant, but it is no longer cheap. Long-term rates have returned to their historical norm, which has exposed a harsh truth: about half of large US companies earn less than their cost of capital (roughly 8% to 11%). Spreading capital across a wide portfolio of businesses is a losing strategy, and higher rates makes it an expensive one.
Spreading capital across a wide portfolio of businesses is a losing strategy, and higher rates makes it an expensive one.
Instead, the current environment favors reducing corporate portfolios to a smaller number of stronger businesses. Firms that pull ahead will concentrate capital on businesses where they have a clear competitive advantage. This is in line with extensive BCG research about how lagging companies can turn themselves around:
- Shrink to a smaller core by shedding underperforming business units.
- Improve returns from that more profitable foundation.
- Grow by reinvesting capital into higher-growth areas and using digital to accelerate development cycles.
Companies that followed this approach delivered roughly 28% in total shareholder return over five years, while those that tried to grow their way out of a weak position created little value. In a scarce, higher-rate market, the same discipline applies to the buildout: win where you are genuinely advantaged rather than chasing every pocket of demand.
Moreover, digital enablement can turbocharge the growth from a profitable core. Embedding AI and connectivity into products does three things at once: it compresses new-product-development cycles, it lifts the productivity of equipment in the field, and it converts an expanding installed base into recurring service, parts, and software revenue that is stickier and more insulated from cycles than equipment sales. For example, Caterpillar connects its 1.5 million installed machines to convert its base into recurring aftermarket and service revenue.
The reindustrialization of America is underway, and its scale is staggering. Massive investments into AI and electrification could draw roughly $6 trillion in investment from 2025 to 2029. Almost every industrial company will be exposed to this shift, either directly or through second-order effects.
But capital is the easy part. To profit from this opportunity, companies need to scale up capacity. They can do so by following the themes laid out above:
- Ride the supercycle by identifying accessible demand.
- Ensure they have the workforce and supply-chain strategies they need for the future.
- Identify new products and solutions that address their customers’ critical supply-chain and labor constraints.
- Localize value chains to get closer to customers.
- Streamline business to focus on a smaller number of stronger positions and then grow from a profitable core.
That’s a challenging endeavor, but the manufacturers that can execute successfully will set themselves up to win the next decade.
