US Growth Driven by "Capital Deepening," Not Hiring
The U.S. economy is seeing strong growth forecasts of up to 3%, but the driver isn't hiring. Analysts on "The Money Lab" podcast argue the economy is undergoing "capital deepening"—shifting from adding workers to giving existing ones better tools like AI and data centers. This is fueling an investment boom in steel, energy, and infrastructure.
This surge in productivity is reminiscent of the late 1990s tech boom, the last period of sustained high productivity growth in the U.S.. Between 1995 and 1999, labor productivity growth nearly doubled to 2.7%, largely driven by private investment in information technology, which accounted for almost half of the increase. That era also saw tech sector wages double over the decade, establishing a significant wage premium over other private sector jobs. Driving today's boom is an unprecedented wave of capital expenditure from technology giants. Four of the largest hyperscalers—Amazon, Google, Meta, and Microsoft—collectively spent over $416 billion on capital expenditures in 2025, a 66% increase from 2024, with plans to spend a collective $700 billion in 2026.. This spending is overwhelmingly directed at building the infrastructure for AI, which S&P Global estimates accounted for 80% of the increase in private domestic demand in the first half of 2025. The investment is not confined to tech. The demand for new data centers is creating a secondary boom in industrial sectors. A single hyperscale data center can require up to 20,000 tons of steel, prompting companies like Nucor and Nippon Steel to invest in producing high-grade steel specifically for these facilities. This has fueled a surge in manufacturing construction spending, which soared 192% between 2020 and 2025, reaching $220 billion. This capital-for-labor substitution is reflected in the cooling job market. While output and productivity are rising, hiring has slowed dramatically. Job growth averaged just 15,000 per month in 2025, and hires rates in key sectors like information and construction have fallen below pre-pandemic levels. This creates a "low-hire, low-fire" environment, where growth is decoupled from adding new employees. The energy required to power this new infrastructure is immense. By 2025, data centers consumed an estimated 4.4% of all U.S. electricity, a figure expected to climb significantly. Projections suggest American data centers will need 35 gigawatts of power by 2030, sparking major new investments in power generation and grid modernization to meet what analysts call "unquenchable" demand. The long-term impact on workers remains a key question. Unlike previous automation waves that affected lower-skilled jobs, AI is increasingly directed at high-skilled tasks. While some studies suggest this could lead to job displacement for 3-6% of the workforce in the coming years, others find that AI-adopting firms grow faster, which can sustain or even expand their overall headcount, particularly in roles that require critical thinking and creativity. This shift may reshape wage patterns. Historically, capital deepening leads to higher wages as labor becomes more productive. However, some economists warn that if the efficiency gains from new technology are not broadly shared through stronger hiring or wage growth for a majority of workers, the U.S. risks a "jobless recovery" that could widen income inequality. The Bureau of Labor Statistics projects that while some analyst roles may see moderated demand, jobs in areas like database architecture and personal finance are expected to grow much faster than average due to AI integration.
