In 1999, internet stocks were minting millionaires. Just months later, most were wiped out.

Investors poured billions into anything with a “.com” in the name. The Nasdaq doubled in a year, but when it crashed, the bottom fell out and more than $5 trillion in market value vanished.

It wasn’t the end of the Internet; it was just the end of the delusion. The same signs are showing with today’s hottest trend: Artificial intelligence (“AI”)

Microsoft (MSFT), Google (GOOGL), Meta (META), and Amazon (AMZN) are expected to spend a combined $750 billion on AI infrastructure between this year and next.

By 2029, total global AI spending could hit $3 trillion, making it one of the fastest capital buildouts in tech history.

That scale alone should raise eyebrows. But history provides clues as to what will come next.

British economist Carlota Perez has studied every major industrial breakthrough of the past 200 years from the steam engine to semiconductors. And each of them followed a familiar arc.

First comes the “installation phase”, where massive investment floods into a new technology. There’s euphoria, overconfidence, and lots of speculative capital.

Since most early bets are based on hype and not utility, the market always crashes. Only then does the “deployment phase” begin. It’s when real productivity gains reshape the economy.

This happened during the dot-com era. Thousands of so-called internet companies went bust. But the internet didn’t die, it matured. Companies like Amazon and Cisco (CSCO) emerged stronger than ever.

AI is now barreling toward that first inflection point.

A recent MIT survey of 3,500 executives found that 95% of companies haven’t seen a return from their AI investments. Most are still testing pilot programs or building tools they don’t yet know how to use..

That doesn’t mean AI is a bust. It means it’s still in a phase where capital flows faster than results.

Meanwhile, the biggest winners are already separating themselves from the pack. OpenAI’s ChatGPT serves over 700 million users a week. Nvidia’s (NVDA) chips are sold out for months.

And companies like Nebius (NBIS) and Oracle (ORCL) announced record orders for AI infrastructure, ramping up their backlogs and resulting in over 60% and 40% rise in stock prices the following day, respectively.

These aren’t speculative plays. They’re becoming essential infrastructure, just like broadband and cloud computing before them.

This shows the importance of picking winners and keeping them close during the massive spending period.

AI is a general-purpose technology like electricity, the engine, or the Internet.

History shows exactly how the trajectory of tech revolutions end. Overinvestment triggers a crash. Weak companies fold. Capital flees the space. But the survivors take everything.

By 2004, five years after the dot-com collapse, Amazon’s stock had recovered and gained over 1,000%. Cisco became the backbone of global internet traffic. Google had just gone public and changed the world of information forever.

The market is headed for the same crossroads with AI.

The hype cycle will eventually pop. Companies with no business model will disappear. But those with scale, infrastructure, and proven demand will enter the golden age of deployment

The key for investors is to stay in the market and own the companies that already have their foot on the gas. Because when the dust settles, those leaders won’t just survive… they’ll dominate

That’s how wealth gets built during tech revolutions. Not by chasing the hype but by holding the winners.

Best regards,

Joel Litman & Rob Spivey
Chief Investment Officer &
Director of Research
at Valens Research

Today’s analysis highlights the same insights we share with our FA Alpha Members. If you want to an get in-depth analysis of market trends and uncover undervalued stocks, become an FA Alpha Member today.