From railroads to GPUs: Why today’s AI investment looks like a wartime economy


While Big Tech’s outsized influence on stock markets is well documented, its role in propping up the broader U.S. economy is less obvious — and increasingly tied to the AI boom.

Driven by artificial intelligence, the largest tech companies have committed hundreds of billions of dollars toward data centers and hyperscaler infrastructure.

Financial disclosures from Amazon, Meta Platforms, Microsoft, and Google-parent Alphabet revealed that they plan to allocate $155 billion this fiscal year alone.

ADVERTISEMENT

According to Deutsche Bank research, without this surge in spending, U.S. GDP growth would be hovering near recession levels.

Deutsche’s data compares quarterly real GDP growth (adjusted for inflation) with and without technology investment. With tech spending included, GDP growth remains positive, though unstable, showing sharp surges during business-cycle upswings.

Stripped of tech investment, however, GDP growth falls sharply and dips below zero in the 2023–2025 period, signaling near-recessionary conditions.

“The AI boom, driven largely by Nvidia and hyper-scaler investments, has been the main force keeping growth positive despite tariff and immigration shocks,” wrote Stock Market News, noting that sustainability is in question as hyperscaler capital expenditures are forecast to peak this year, leaving the economy without a clear momentum driver.

The dangers of over-reliance on tech

AI development is widely seen as a net positive, driving productivity gains and enhancing business profitability. Yet the economy’s heavy dependence on tech spending to sustain growth and support financial markets leaves the broader system vulnerable to a downturn in a single sector.

As Deutsche Bank’s research highlighted, technology capital expenditures are highly cyclical and subject to financing costs, regulatory headwinds, overcapacity, and shifting consumer demand.

While AI and cloud infrastructure could deliver lasting productivity gains, history shows that transformative technologies take time to ripple across industries and meaningfully lift the wider economy.

ADVERTISEMENT

To illustrate how unsustainable current trends may be, Derek Thompson of The Atlantic noted that AI-related capital expenditures could already account for a bigger share of GDP than any technology investment since the railroads of the 1880s.

“Basically, it’s a mini-wartime economy, but the guns are chips and the tanks are databases,” he wrote.

AI engineer and entrepreneur Rohan Paul has also flagged an overlooked risk: while AI data centers are absorbing enormous amounts of capital, the infrastructure is depreciating at an estimated $40 billion annually — roughly double current revenue trends.

“To earn a normal return at this scale, United States data centers would need about $480 billion revenue in 2025, far above current run rates,” Paul said.


ADVERTISEMENT

Leave a Reply

Your email address will not be published. Required fields are markedmarked