Understanding America’s Uneasy Dance with Technology and Industry

In recent years, the  United States government  has attempted to reverse decades of  deindustrialization  through tariffs on imports, particularly from China. However, a new wave of investment in  artificial intelligence (AI)  is unfolding, echoing the very patterns that led to the decline of American industry in the 1990s. History often repeats itself, but this time, we have a clear understanding of the potential consequences.

Why this matters: Derek Thompson, an insightful business reporter at The Atlantic, has pinpointed a critical trend in the landscape of American industrial decline. His analysis suggests that while China may have stolen jobs from American workers, it was also the  abandonment of American capital  that played a significant role in this gradual decline. As factories closed, the investment landscape shifted dramatically.

In a captivating interview with investor Paul Kedrosky for his podcast, *Plain English*, Thompson outlines his thesis:

  1. The 1990s saw a  massive deployment  of the Internet and telecommunications, drawing substantial amounts of investment.
  2. This outflow of capital left small manufacturers grappling with costly financing just as China’s entry into the  World Trade Organization (WTO)  began to remove trade barriers.

These shifts did not stem from mere coincidence; they were a chain reaction— cause and effect  interlinked.

The context: This year, technology companies are projected to invest approximately  $400 billion  in infrastructure for AI. To provide a frame of reference, this amount almost equates to the cost of the Apollo program that successfully landed humans on the Moon, which totaled around  $300 billion  when adjusted for inflation. However, while the Apollo mission spanned a decade, the current AI investments are being made in just one year.

So far, data centers have contributed to about  half of the U.S. economic growth  recorded in the first six months of 2025. Projections indicate that global investment in data centers could surpass  $500 billion annually  by 2026 and 2027.

In contrast, American consumers currently spend just  $12 billion  a year on AI services, revealing a stark  disparity  between investment and returns.

The panoramic view: This situation exhibits structural issues. If an investment fund possesses  $500 billion , fund managers face two choices:

  1. Distribute that bulk among a hundred small factories needing around five million each.
  2. Opt to issue ten lucrative checks worth  $50 billion  to promising AI projects.

The choice is evident: the first option involves managing a proliferation of companies, while the second requires far fewer meetings and oversight.

  • For manufacturers eager to bring production back to the U.S., access to financing is painfully expensive.
  • Banks look at these manufacturers’ projects in contrast to the alluring returns promised by AI investments.

This puts local production at a distinct disadvantage.

The irony: Former President Trump constructed his economic strategies around tariffs designed to encourage American manufacturing. Yet, investment in AI simultaneously drives up the expenses associated with local production—exactly the outcome the tariffs aim to counter.

  • Tariffs inflate the cost of imports from China.
  • AI increases financing costs for domestic manufacturing.

The resulting effect may prove negligible for the industry while escalating prices for consumers.

The figures: Constructing a modern data center typically entails:

  •  Sixty percent  of the budget allocated to  NVIDIA chips .
  • The remaining budget dispersed among refrigeration, electricity, and construction.
  • The actual physical building costs are relatively inexpensive compared to technology investments.

Geography plays a significant role, as Northern Virginia has captured a large portion of this investment. Regions that were once rural are now dotted with industrial facilities operating around the clock.

Yes, but: An option now exists that was unavailable during the 1990s: establishing data centers outside the United States. Nations like India and various Middle Eastern countries are receiving  substantial investments  due to cheaper electricity and fewer local disturbances. However, this trend exacerbates the existing problem. As capital flows to overseas data centers, even less investment is left for American manufacturers.

Between the lines: Kedrosky likens the current situation to a  Death Star absorbing capital . In the 1990s, that gravitational force was the Internet; today, it is AI. Unfortunately, factories have once again become collateral damage. The crucial difference, however, is that we now recognize the potential repercussions of these shifts.

The United States stands at a notable crossroads. With the lessons of the past illuminating the decisions of the present, the nation must determine whether it will encourage a domestically focused strategy that fosters sustainable industrial growth or continue down a path of reliance on technologies that threaten its manufacturing base. Understanding this delicate balance will be vital to shaping a successful economic future.



General News – 2