Silicon Valley Is Hunting for Its Next $1 Billion Bargain

Listen to the audio version of this article (generated by AI).

In May 2023, Spark Capital made the biggest investment in its history.

The venture-capital firm wrote an initial $75 million check to help fund Anthropic, then a little-known artificial intelligence startup trying to compete with OpenAI. Anthropic had no stock symbol, and it generated very little revenue. There was no Wall Street research report telling investors what it was worth.

Three years later, Spark’s stake is estimated to be worth $7 billion on paper. 

Spark Capital only needed one Anthropic. History suggests investors often don’t need many, either.

I keep coming back to that story because I think it marks the beginning of the current AI investment cycle. Investors like Spark provided the capital that helped companies like Anthropic grow into some of the most valuable businesses in the world.

But that cycle is already changing.

AI’s Builders Are Becoming Buyers

The companies that were startups just a few years ago are fast becoming giants themselves: Anthropic, OpenAI, and others. Along with Microsoft Corp. (MSFT), Alphabet Inc. (GOOG), Amazon.com Inc. (AMZN), Meta Platforms Inc. (META), and Nvidia Corp. (NVDA), they’re raising – and spending – enormous amounts of money, and racing to build what they believe will become the next great computing platform.

Sooner or later, though, they’ll find they can’t invent everything themselves. No company — not even the AI labs valued near $1 trillion — can hire every brilliant engineer or invent every breakthrough first.

Long story short, I think we’re entering the next phase of the AI boom.

The first phase rewarded investors who recognized that AI infrastructure – chips, memory, networking, power, and data centers – would become essential. I still believe many of those companies have room to run.

But the next phase could look very different.

Instead of asking, “Which companies will build AI?” investors may soon need to ask, “Which companies will AI’s biggest players decide they have to own?”

If I’m right, answering that one question could make the difference between simply participating in the AI Revolution… and getting there before the rest of Wall Street catches on.

That’s what I want to show you today.

First, I’ll explain why I think the AI arms race is entering a new phase. 

Then I’ll show you how Silicon Valley’s biggest companies tip their hands long before they announce their next blockbuster acquisition. 

Finally, I’ll explain the framework I’ve started using to identify the kinds of companies I believe could become tomorrow’s biggest winners, whether they eventually go public or get bought first.

Let’s start by following the money…

Follow the $700 Billion AI Arms Race

One of the first rules I learned as an investor is that money leaves clues. When hundreds of billions of dollars begin flowing in the same direction, I pay attention.

Right now, the money isn’t just flowing into Nvidia.

It’s flowing into the entire AI ecosystem.

Follow the money. 

Amazon alone expects to pour roughly $200 billion into capital projects this year. Microsoft and Alphabet are each planning about $190 billion. Meta could spend another $135 billion. 

That’s roughly $700 billion in a single year, or about $2 billion every single day. 

That’s an arms race. 

Meanwhile, Anthropic’s latest funding valued the company at around $965 billion. And OpenAI is said to be worth roughly $852 billion.

But even when companies have that kind of money, they can’t invent everything themselves. Call it Silicon Valley’s dirty little secret: It’s often faster to buy innovation than build it yourself.

AI Giants Will Buy What They Cannot Build Fast Enough

Think about what these companies are trying to accomplish.

The companies leading AI genuinely believe they’re building the next computing platform. When the stakes get that high, companies stop asking, “Can we build this?” and  start asking, “Who already has?”

We’ve seen this movie before.

In 2012, Facebook paid $1 billion for Instagram. At the time, it looked ridiculous.

Instagram had just 13 employees. It wasn’t making money. Most people thought Mark Zuckerberg had wildly overpaid for a photo-sharing app that let people put vintage-looking filters on pictures of their lunch.

Turns out, Zuckerberg got the bargain of the century. Instagram has since become one of Meta’s most valuable businesses. Last year, Meta estimated Instagram’s brand value at over $70 billion, and said it generates nearly $67 billion in annual revenue. 

But that’s not what I find most interesting.

The biggest winners weren’t the people who bought Meta stock after the Instagram deal was announced. They were the people who already owned a piece of Instagram before Zuckerberg came calling.

And this pattern continued throughout the last few tech booms:

Google bought Android before smartphones became ubiquitous and picked up YouTube before online video dominated media. 

Facebook bought Instagram and WhatsApp before those businesses reached their full potential. 

Microsoft acquired GitHub as software development became increasingly collaborative and cloud-based. 

Notice the pattern. None of these companies were acquired because the buyers were running out of money. They were acquired because the buyers were running out of time.

I think AI is setting up a similar dynamic, only on a much bigger scale.