America’s AI investment boom is real. However, this surge has yet to result in the kind of industrial revolution that many had hoped for.
To be sure, the United States is experiencing a genuine boom in AI investment. The hyperscalers accelerated their R&D and capital investment 50-fold over the past two decades to $750 billion in 2025 compared to $15 billion in 2005. By the end of 2026, total investment by these hyperscalers could approach $1 trillion.
But here’s the catch: despite that record level of spending, productive investment as a share of U.S. GDP has barely budged. That measure matters because it tracks spending on the economy’s productive assets—the factories, equipment, infrastructure, and intellectual property that underpin future productivity and competitiveness. The AI boom is real, but it has yet to reshape investment across the broader industrial economy.
This disconnect matters because productive investment is a leading indicator of competitiveness and of where production, jobs, and growth will occur. While the United States has outperformed most advanced economies on investment since the global financial crisis, China is adding roughly $4.4 trillion in net productive assets annually, roughly four times the equivalent amount in the United States.
We estimate that addressing the most critical U.S. import dependencies could require on the order of $2 trillion in additional manufacturing investment or about 6% of GDP. At the end of 2025, investment in factory structures fell 6% after peaking in 2024. Meanwhile, investment in general industrial equipment was essentially flat, although there was a slight uptick in machinery and equipment investment in the first quarter of this year. The reshoring momentum that started in 2022 has plateaued in the numbers, and any recent announcements will take time to translate into construction and development.
One clear challenge to sparking a U.S. industrial renaissance: it’s expensive to make in America. Across most steps of the production process—construction, labor, materials, equipment, and time to market—the United States is a costly place to invest. Excluding any subsidies, the all-in costs to build products like semiconductors and pharmaceuticals are roughly 40% and 60% higher, respectively, than in the most competitive locations, while the cost of developing a new antibody medicine is 2.7 times as expensive compared to China. Two factors constitute the bulk of the cost gap. The first is more costly and slower capex delivery. U.S. construction costs are about double what they are in Asia, and construction times can be twice as long: recent nuclear projects have taken up to a decade to complete compared to six years in China. Second, labor costs are two to five times more than in China or Taiwan, a difference that used to be offset by productivity differences. But in like-for-like industrial settings, productivity differences have all but vanished. In advanced fabs, for instance, Taiwanese engineers produce about a quarter more per worker than in the United States, where wages are more than 2.7 times as high.
To close such gaps, companies hoping to build at home could start by using modular, off-site methods that can cut project timelines by half and capital costs by 10 to 20%, and deploying technology, collaborative contracting, and more to lower construction costs. Also AI- and robot-first operating models can help employers transform labor productivity. Our analysis found such steps could close half to two-thirds of the U.S. cost gap.
Where cost competitiveness isn’t possible, companies can compete on service quality, brand, customer proximity, and innovation. Complex drug therapies, for example, command premium margins and a decade or more of effective commercial exclusivity. Performance and trust can sustain premium prices. Increasingly, unrestricted access to the U.S. market also matters.
Policymakers face their own challenges. They cannot protect, nurture, ringfence, or subsidize every industry. Instead, they can support industries that can solve America’s so-called “Achilles heels,” the roughly 25% of imported manufactured goods that are critical to national security, exposed to supply concentration, and derived from geopolitically distant trading partners. The scale of intervention required, whether selective trade measures, financial support, industrial policy, or other measures, is substantial. The task is about triage, deciding which industries justify a scale of intervention that would change the playing field, starting with the 25% of imported manufactured goods in which dependencies are most pronounced. Policymakers will also want to work to address existing skews in the international trading system.
The U.S. technology boom shows the country can mobilize capital at breathtaking speed. Kicking off an industrial renaissance raises a harder question: Is the United States willing to change how it builds industry and absorb higher costs when doing so isn’t enough?
The post Why America’s AI Boom Isn’t an Industrial Boom appeared first on TIME.
