Serial Internet Wave Rider

Jeremy Allaire graduated from Macalester College in St. Paul in 1993 with a degree in political science and philosophy and no ability to write a line of production code, a limit he acknowledges freely. The first graphical web browser reached the public that same year, and Allaire declared himself an internet consultant on the spot, three years into an obsession with a network that was still mostly gopher menus, over his father's considerable concern.¹ He was convinced this obscure academic network was "going to reshape how communications and media and software works."¹ He was the better part of a decade early.

He would be the better part of a decade early twice more. In 2003, having just put video into Flash as Macromedia's chief technology officer and watched it reach 98 percent of the world's computers within a year, he left to build Brightcove, on the theory that television itself was about to move onto the internet.¹ In 2013, immersed in the strange new world of cryptocurrency blockchains, he decided dollars were about to move onto the internet the same way, stepped down from the chief executive seat of a company he had already taken public, and started Circle.¹ Web applications, internet video, internet money. Three companies, three IPOs, each begun eight to ten years before the crowd arrived.¹

That is the whole archetype, and it is stranger than it sounds. Allaire cannot build any of the things he sees coming, which in his case is a biographical fact rather than a requirement of the plate. What he can do is stand at the edge of a technology before it has a market, describe, correctly, the utility it will become, and then build the layer underneath it and wait for the world to need it. Under all three companies is that one move. Find a technology that works technically before it works commercially. Study it until the utility comes into view. Build the infrastructure layer that thousands of applications will sit on rather than any single application. Set the clock in decades and hold through the years when almost nobody agrees. ColdFusion was built before enterprises demanded server-side web apps, Brightcove before media companies accepted that TV was moving online, Circle before any regulatory framework for stablecoins existed.¹ The judgment was identical every time: this technology makes something possible that will be widely adopted, and I will build the layer for it now.

The study is the mechanism, not the degree

The method is close study of a machine he does not operate. "I need to understand these things really really deeply. And that doesn't mean you would trust me to write software, you wouldn't, but I need to know enough that I can work really closely with software architects, product designers, engineers, product leaders."¹ The whole method is in that sentence, and the load-bearing half is the second clause, not the first. He gets close enough to the architects to translate what a technology makes possible into what it should become.¹ This is why the plate is teachable in a way a natural-born builder's is not: what transfers is a discipline of study, and a discipline of study does not care whether the student can type.

He builds the layer beneath the applications, never the application. ColdFusion was a framework other people built web apps on. Brightcove was infrastructure media companies published video through. Circle is a monetary network any developer can move dollars across without ever calling Circle.¹ Each time the instinct is to be the platform for thousands of applications rather than one of them, a slower and much larger way to win, because a platform compounds on other people's work. And he has watched twice what the wave does to incumbents: the phone call became FaceTime, the cable bundle became streaming, and in his telling the bank wire is next in that queue, with banks surviving the way the phone companies did, by building on the new rails.¹

Each wave teaches him to read the next. The recognition is cumulative. The developer-platform network effects he learned running ColdFusion are what let him see Circle's moat; dragging a non-technical industry onto internet rails at Brightcove is what prepared him to do it to money.¹ He did not ride three unrelated waves. He used each as a lens on the next, which is why the third bet looked reckless from outside and close to obvious from where he stood.

He lobbies for the rules that will bind him, from day one. When Circle launched in 2013, its founder was advocating for a stablecoin framework while unable to open a bank account, hire an auditor, or buy insurance, because a search for his technology returned Silk Road and criminals, and his own family office passed on the round.¹ Hundreds of stablecoins have launched since, and are, in the words of one of his colleagues, "stable in name only."¹ The Wave Rider will live in the legal gray zone for years while campaigning for the rules that end it, because the rules that bind him also legitimize the space and clear out the operators who were never serious. That is the long outcome bought at the price of the near-term crisis, held the way he holds everything: for years.

He argues the moat is political philosophy expressed as architecture. Allaire frames the open, permissionless, decentralized DNA of the internet and of blockchains as liberal enlightenment ideals written into protocol, and argues this is exactly why no Chinese equivalent of Circle exists: a society built around centralized control does not produce permissionless money, however much capital it commands.¹ Any analyst can price the network effects. The deeper claim is that the values are encoded in the wiring, and a state cannot fork values.

Every tell on that list is a reading, not a build.

The third time, he was running the play on purpose

This archetype is recognized, which is its own way of arriving. Chosen archetypes read like strategy and stumbled ones read like scar tissue; Allaire's arrived as a shape he kept noticing in the world until he noticed that the person seeing it first was always him.

The first time, in 1993, there was no pattern, only the obsession with a pre-commercial network. Repetition made it legible. By 2003 he had done it once. By 2013 he was watching deliberately, and watching for the right thing: what a new technology quietly made possible, years before any market asked for it. His own account of the recurring insight is almost bored. The internet was doing something new to an old domain again, communications first, then media, then money.¹

The fuel underneath is conviction, and it has to be, because there is no near-term reward to run on a decade out from the market. What sustains the years of disbelief is the sentence he hands to everyone he recruits: "This is like a 10 to 20 year thing."¹ He said it to the first employees and the first board members. Twelve and a half years later the sentence has survived a 25-times-oversubscribed IPO, $25 billion of demand chasing a billion-dollar offering priced at $31 after multiple upsizes, and a market capitalization around $20 billion, and his read is unchanged: Circle is "actually an early stage company" with another ten to twenty years in front of it.¹ Optimism would be cheaper. The sentence is a piece of engineering, the framing required to raise money and hire people for a project that will outlast the fund cycle of nearly everyone writing the checks.

Most of this is copyable. The read is not.

Watch enough founders and the shapes start repeating in different bodies; most entries in this taxonomy are temperaments wearing a strategy, and a temperament cannot be installed. This entry is closer to a checklist, and Allaire is the reason the taxonomy can admit learnable entries at all. Being a natural builder is not a decision. Doing what Allaire does is: picking a technology that works before it sells, studying it until the utility it will become can be described, building the infrastructure rather than the flashy application, and setting the horizon in decades so the early years of disbelief do not break the builder. The mechanics are teachable, and a humanities graduate ran them three times, which is the proof.¹

What does not transfer is the thing the mechanics depend on. The wave-identification instinct either fires correctly or it does not, and there is no way to check the work in advance. The infrastructure discipline, the regulatory patience, the decade-long framing: all of it is worthless if the called wave never breaks. The recipe transfers completely and guarantees nothing, because the one step that matters is the one nobody can teach.

For the first decade, right and wrong look identical

Which is the failure mode, stated plainly. The pattern looks obvious in retrospect and is close to unfalsifiable in advance.¹ From here the sequence reads like destiny, web apps and internet video and internet money, of course. In 1993 and 2003 and 2013, each call looked to serious people like a bet that could not be settled for a decade. The instinct that produced three correct reads is indistinguishable, on the morning it fires, from the instinct that produces a patient, well-capitalized, decade-long march into water that never rises. The Wave Rider who is early and the Wave Rider who is simply wrong hold the same conviction, recruit with the same speech, and report the same absence of traction, for years.

Only the water tells them apart, and it takes its time.

There is a partial defense, and Allaire leans on it: he bets that a layer will be needed, never on which applications will win, which hedges everything except the thing that matters. Against being wrong about the wave itself there is no defense at all.

Three waves, ten years apart

Set the Wave Rider beside the wave recognizer and the honest difference is not the number of waves, which is arithmetic nobody can run on a founder mid-career. It is what the edge is made of. CZ's edge is an operating skill that does not travel, and he says so plainly. Asked where he would build if he were starting from nothing: "I have to pick something that I know how to do. And my experience only leads me to building an exchange."² On the boundary of it: "If I want to lead an AI team, that would probably be a disaster. So I have to stick to what I know best."² Allaire's edge is a way of studying a technology until its utility comes into view, and that method has been pointed at web applications, at video, and at money without changing shape. A skill fixes a founder to one wave. A method follows the water. Masayoshi Son is the third case, building the capital layer that funds everyone else's infrastructure; Allaire builds the infrastructure himself, one layer per wave, and then stands at it, waiting.¹

The sequence itself is regular. In 1993 it was web applications, called while the internet was still gopher menus. In 2003 it was internet television, called the year video first played smoothly in a browser. In 2013 it was internet dollars, called while the industry's search results still said Silk Road.¹ Eight to ten years ahead of the market every time, and ten years apart, to the year.

The 1993 call was made on a conviction he could not yet justify. He was not smarter than his classmates about software; he could not write any. He was earlier. Three times the water came in behind him, on roughly the schedule he named, and on the morning of each call there was no way for anyone, including him, to know.

Allaire has already named his own candidate for a fourth wave: dollars moving between AI agents "in a fraction of a second, at a fraction of a cent," a use he says is already live, and he has given it his usual horizon, another ten to twenty years.¹ Whether that read is correct cannot be known in advance, which is the archetype's defining condition. Three correct calls establish the pattern. They do not settle the fourth.