Governments Pour Billions into Chips and AI Infrastructure to Fuel Arms Race

One of the most critical resources in 2025 is compute power. Chips and the data centers that house them have become the 21st-century equivalent of oil refineries and power plants, and governments are increasingly treating them as such.

Policymakers around the globe, from Washington to London to Beijing, are pouring billions into semiconductors and cloud infrastructure, not only to gain an economic edge but also to lead in artificial intelligence (AI).

Just look at OpenAI. Earlier this month, the company signed one of the largest cloud contracts in history with Oracle—$300 billion worth of computing power spread across roughly five years. The deal will require 4.5 gigawatts of power capacity, or the equivalent of more than two Hoover Dams, if you can believe it.

OpenAI’s spending spree has been the rising tide that’s lifted all boats in tech. Since ChatGPT burst onto the scene in late 2022, the market caps of Nvidia, Microsoft, Oracle and Broadcom have swelled by a staggering $8 trillion. That, of course, is a big reason why the Nasdaq and S&P 500 have continued to hit record highs this year.

Billions Flowing into Data Centers

The AI revolution has ignited an unprecedented building spree in U.S. data centers. According to Bank of America, construction spending hit an all-time high of $40 billion in June alone, representing a 30% increase from the year before. That’s on top of a 50% surge in 2024.

Washington isn’t sitting on the sidelines. Just ask BAE Systems. With funding from the CHIPS Act, the London-based defense contractor is currently modernizing its 110,000-square-foot Microelectronics Center in New Hampshire. The facility is one of the few domestic manufacturers specializing in military-grade semiconductors.

Governments Supporting AI Technology

Last month, Intel made a historic deal with the Trump administration. The government announced it would take an $8.9 billion equity stake in Intel, in addition to billions in CHIPS Act grants. (And just this week, Nvidia said it would be investing $5 billion in the struggling tech firm.)

I believe the message is loud and clear: Semiconductors are strategic assets like oil and critical metals, and Washington is willing to invest taxpayer money to support them.

Across the pond, the UK is also ramping up AI investment. Microsoft announced plans to invest $30 billion by 2028 to build the country’s largest supercomputer, equipped with more than 23,000 Nvidia GPUs. Google, Nvidia, OpenAI and Salesforce are also pledging billions. All combined, tech giants are pouring more than $40 billion into the UK’s AI infrastructure.

Nvidia, Broadcom and AMD Fight for Industry Dominance

So who’s making the chips behind the revolution?

Nvidia remains the undisputed king, according to Morningstar. The financial services firm projects that AI chip revenue will quadruple over the next several years, with Nvidia leading the way.

But they’re not alone. Broadcom is carving out a strong second position, while AMD is battling for share in general-purpose GPUs.

Biblical Amounts of Energy Will Be Needed

As impressive as these chips are, they demand enormous amounts of energy. The Department of Energy predicts AI-specific electricity use will grow at a mind-boggling 33% each year, far exceeding past norms.

The entire AI ecosystem relies on whether we can build and deliver enough electricity to feed it, which is why I think investors should be paying as much attention to power grids and cooling systems as they are to chip stocks. I said as much in an article last month.

High-Tech Military Applications

The connection between semiconductors and defense is only increasing. As I’ve been writing about recently, global defense budgets are expanding rapidly. PwC estimates that military spending will grow from just under $3 trillion in 2024 to potentially $4 trillion by 2030, with an increasing portion of that allocated to high-tech systems like drones, satellites, autonomous ships and AI-enabled fighter jets.

Is AI Contributing to the Labor Market Slowdown?

Overlay all of this with the macro picture. The labor market appears to be softening. Job openings are declining, and small businesses are cutting back on wage hikes. CEOs across industries are talking about white-collar automation.

In my mind, this creates a paradox: AI is being sold as a productivity booster at a time when the economy is moderating. Historically, periods of weak labor demand have often coincided with waves of automation. It’s no coincidence that businesses are investing so heavily in AI just as hiring slows.

That said, I would urge investors to recognize that compute is the new energy. Nations are stockpiling it, and companies are monetizing it. Just as oil defined the 20th century, I believe compute will define the 21st. The question isn’t whether to get exposure, but how.

Interested in investment opportunities in the fast-growing semiconductor, data center and AI space? Email us at info@usfunds.com with the subject line CHIPS.

—

Originally Posted on September 22, 2025 – Governments Pour Billions into Chips and AI Infrastructure to Fuel Arms Race