At Founders Fund, I hear dozens of pitches for new companies every month, and for the past year or so — regardless of who the founders are, or what product they’re pitching — someone will say “artificial intelligence” within the first five minutes of our meeting.

AI is a transformative technology (and we definitely don’t need another think piece explaining why). There will be AI companies that upgrade daily life to at least the same degree as fiber optic cable or smartphones. But, as a buzzword, AI’s sudden ubiquity is not purely a function of its business potential.

The average investor does not spend all day simply searching for the best new company; they spend all day searching for the best new company in a socially approved technology “space.”  Twenty years ago, that space was e-commerce, the “dot.com”-ing of brick-and-mortar stores. Fifteen years ago, “social-mobile-local.” Ten years ago, the “sharing economy.” Three years ago, gaming and crypto. Now, it’s artificial intelligence.

This herd behavior is not harmless.

A hyped venture space only exists after the winning company in that space has already been created, discovered, funded, and become popular. The very existence of the space proves the search for a winning investment is already over. This is the reality of venture investing, and the herd behavior operating in open defiance of this reality is necessarily limiting fund returns while denying capital to promising companies simply because they aren’t on trend.

Consider the hype now emerging around defense technology, which is primarily sparked by the success of Anduril, a company I co-founded that develops surveillance, weapons, and intelligence systems for our government.

When we started the company in 2017, defense technology could not have been more unpopular. Prior to Anduril, one of my co-founders, Palmer Luckey, started the virtual reality headset company Oculus at 16 years old and sold it to Meta for $2 billion five years later. This was Steve Jobs-level precociousness, and yet Anduril was still met with derision in the media and skepticism in pitch meetings.

The reasons were endless. We were warned it would be impossible to navigate a sclerotic government contracting process dominated by a small group of massive corporations with vast lobbying muscle. Gen X and millennial investors had no memory of great power conflicts and simply assumed the era of serious state-on-state violence was over.

Some prospective investors outright admitted they were worried an association with the armed forces would compromise their business prospects in China. And as evidenced by Google’s abrupt decision to cancel its AI drone contract with the Pentagon after about a dozen employees resigned in protest about it, Silicon Valley is skittish about technology projects that could become a target in the culture wars. Several brand-name investors were wary of having any association with Palmer, who had been unceremoniously ejected from Facebook for the thoughtcrime of supporting Donald Trump.

Today, Anduril is valued at more than $10 billion, and the hype has followed. In 2014, just $200 million in venture investment flowed into defense technology; last year, that figure topped $6 billion. Investors see our success as proof the defense industry can generate serious returns, and now they’re hoping to get a piece of the next Anduril. But, in all likelihood, another defense tech company operating at Anduril's scale won’t come out of this hype cycle.

Defense, like all technology sectors, is ruled by the power law: a single, prime mover will claim the vast majority of profits even after the sector has reached maturity. The company that carves out a new investment category usually stays on top. Network effects provide a compounding advantage that make it very difficult to dislodge the first mover. This dynamic has profound implications for venture investing. It means that by the time an industry is the focus of the hype cycle, it’s already too late — the chance to generate substantial returns has already passed. Investors should be trying to get a piece of companies that create, not follow, hype.

Consider Uber. Fifteen years ago, Travis Kalanick and a couple friends, frustrated with San Francisco’s grimy-but-pricey taxi services, built a mobile app to hire luxury cars. That’s the birth of Uber, which made on-demand ride-sharing, once unfathomable, a default feature of modern living.

Uber wasn’t just a unicorn, it was the creation of a new investment category. And its success sparked a predictable venture hype cycle, with billions of dollars funding dozens of new competitors. Expanding required Uber to fight through the regulatory system of essentially every major metropolitan region in America, a morass aggressively guarded by taxi associations watching their cartel power disappear into dust.

And yet, fifteen years later, Uber is still the dominant player. At $162 billion, its market cap is triple that of the next ten ride-sharing companies combined.

This is the power law, and it’s remarkably consistent across product categories. At $47 billion, Coinbase’s market cap is more than double the next ten crypto companies combined. At $180 billion, SpaceX’s market cap exceeds that of the next twenty space companies combined. And at $1.2 trillion, Meta’s market cap is more than double the next dozen social media companies combined.

In venture investing, the power law companies are what matter. These are the investments that return the fund. Venture firms that blindly follow the herd and only cut checks to sectors in the hype cycle are acting in open defiance of their own industry’s iron law. Once a category becomes mature enough to attract hype, it’s already too late for new investment. The power law firms will continue to claim most of the profits.

Following the herd also runs the risk of ignoring genius founders who aren’t on trend. There is undoubtedly Palmer Lucky-level talent being ignored by investors today because it’s not focused on AI or defense. Hype culture slows the rate of innovation by denying capital to entrepreneurs pursuing genuinely world-changing technologies.

There’s also the inconvenient fact that’s almost never mentioned during any of venture capital’s endless Zoom calls, Sand Hill power lunches, or glitzy conferences: the hype cycle hobbles founders. Hype-driven valuations can be a burden, damaging the long-term prospects of the company.

That’s the dynamic that drove what’s probably the most famous startup flame-out of the last decade: WeWork. The company started with a genuinely innovative strategy of investing heavily in design, culture, and amenities, clearly distinguishing itself from the drab cubicle farms that had long defined co-working. And that brand edge allowed WeWork to attract the self-employed and small businesses rapidly proliferating in the burgeoning digital economy. It was a solid, if not exactly world changing, business idea.

But solid is insufficient in a hype cycle. It’s not enough to just be an innovative co-working company. When WeWork found itself in the red-hot center of the co-working investment boom, it came under tremendous pressure to become something more lucrative. Softbank CEO Masayoshi Son famously badgered WeWork CEO Adam Neumman — a man no one would accuse of being short on ambition and who was already securing big city leases at an unprecedented rate at peak market prices — to get even more aggressive with growth.

As a result, WeWork locked itself into an unsustainable sub-leasing structure. Then, COVID pushed the company into calamity, as knowledge workers migrated to work-from-home en masse and never came back. Once valued at $47 billion, WeWork filed for bankruptcy last November.

Instead of hype cycle investing VCs should bet on founders, not sectors. That necessarily means rejecting hype-driven consensus. Generating substantial venture returns requires breaking from the herd and betting on companies that launch, not follow, hype cycles.

Instead of evaluating pitches on a relative standard — “How does this particular AI start-up compare to all the other AI start-ups?” — VCs should use an absolute standard. Relative wins are effectively irrelevant to the returns of any large fund. The real question to ask is: “Could this company become a category-creating monopoly? Could it come out on top of the power law?” Because those are the only companies that matter.

Betting on founders fulfills the proper function of venture: to generate returns for investors by providing resources to talented founders to create products and services that upgrade reality. And it’s the category-creating — not hype-following — companies that do that.

—Trae Stephens