Remember when Spark Capital dropped $75 million on Anthropic back in 2023? Most people thought they were nuts. Today, that stake is worth about $7 billion on paper. Not a bad return for betting on a startup nobody knew existed.
Here’s the thing though: that era is basically over.
The AI startups that were scrappy underdogs three years ago—Anthropic, OpenAI, and friends—are now the giants. And like all giants, they’re about to get hungry. Not for more funding, but for acquisitions.
**The $700 Billion Shopping Spree**
Let’s talk about the money flowing around right now. Amazon’s planning to drop roughly $200 billion on capital projects this year. Microsoft and Alphabet? About $190 billion each. Meta’s throwing in another $135 billion. That’s $700 billion annually—roughly $2 billion every single day—just on infrastructure and tech.
When you’re spending that kind of cash, you eventually realize you can’t build everything yourself. It’s faster to buy innovation than invent it from scratch. That’s Silicon Valley’s dirty little secret.
**History Rhymes (Again)**
We’ve seen this movie before. Facebook paid $1 billion for Instagram in 2012. Everyone called Zuckerberg crazy. Instagram had 13 employees and wasn’t making money. Turns out, he got the deal of the century—Instagram now generates nearly $67 billion in annual revenue for Meta.
Google grabbed Android and YouTube before they dominated their markets. Microsoft scooped up GitHub as software development went cloud-first. The pattern’s always the same: these companies weren’t acquired because the buyers were broke. They were acquired because the buyers were running out of time.
AI is setting up for something similar, except on steroids.
**The IPO Trap**
Here’s what most investors miss: the stock market isn’t where great companies begin anymore. It’s where they finish their first chapter.
Look at SpaceX’s recent IPO. Millions of retail investors finally got their shot to buy in. But by then, SpaceX had already spent years building rockets, launching satellites, and creating massive value. Within weeks, the stock tanked below the IPO price. The real winners? The people who got in years earlier at private valuations.
Two people can believe equally in the same company and walk away with completely different results.
**The Framework**
So here’s the question that matters: Which companies will become so important that an AI giant decides it can’t afford to compete against them?
Some will become the next generation of AI leaders. Others will get acquired long before they ever ring the opening bell on Wall Street. Either path can create enormous value—if you spot them early.
The trick is following the money. Look at where Silicon Valley’s investing. Identify what capabilities the big AI players still lack. Find the smaller businesses solving problems the giants might eventually decide they need to own.
**The Bottom Line**
During technological revolutions, headlines focus on the giants. But the biggest fortunes? They’re usually made one or two layers beneath them.
The first phase of AI rewarded companies building the infrastructure. The next phase could reward the companies those builders decide they have to own. And that’s where the real opportunity might be hiding.
