Hayden Adams and the market that began as homework

Hayden Adams had helped build an exchange, arranged its security checks, and traveled to Prague to launch it. Then he spent about an hour on a beanbag, rewriting a tweet. It was November 2, 2018. A friend, Ashleigh Schapp, stopped to help him review the words. The machinery was ready. Announcing it to people still required another person.

That small pause belongs in the history of Uniswap. Adams is its creator and the founder and CEO of Uniswap Labs, the New York software company. His work concerns markets that can run through code, without a company taking custody of every trader’s money. His career, meanwhile, has depended on a rather old technology: friends persuading friends to try something.

The engineer with an unexpected opening

Adams studied mechanical engineering at Stony Brook University, graduating in 2016. At Siemens, his work involved computational fluid dynamics, the study of how fluids and heat behave. It was a plausible beginning for an engineering graduate. In a later interview, he described being distinctly unenthusiastic about it. He also thought he probably would have stayed had the decision remained his.

On July 6, 2017, he was laid off. His college friend Karl Floersch, then at the Ethereum Foundation, responded with congratulations and an improbable career suggestion. Adams should become an Ethereum developer. His objection was reasonable: he had very little programming experience. Floersch’s answer was that smart contracts were new enough for a beginner to catch up.

There is a useful distinction here between acquiring a new skill and finding a reason to use it. Adams learned Solidity and JavaScript, but exercises could take him only so far. Floersch pointed him toward Vitalik Buterin’s writing about automated market makers. Here was a project with an intelligible engineering problem: could software let people exchange two assets according to a rule?

“Don’t I need to like know how to code?”

Hayden Adams, recalling his first conversation about changing careers

A market with an equation at its center

A conventional exchange collects orders from people who want to buy and sell. An automated market maker offers a different arrangement. Participants supply assets to a shared pool. Traders exchange against that pool, and a formula determines the terms. Liquidity providers receive trading fees for making their assets available. The pool can keep operating without waiting for a particular buyer and seller to agree.

The early Uniswap design used the constant-product rule, x × y = k. The two quantities of tokens multiply to a constant, before accounting for fees. Take one token out and more of the other must go in. As the available balance shrinks, buying another unit becomes more expensive. A small formula does considerable work, though it cannot make a thinly funded market behave like a deep one.

Adams did not claim to have invented the underlying idea. He credited Gnosis’s Alan Lu for conceiving constant-product market makers on Ethereum and Buterin for bringing the approach to public attention. Adams’s contribution was a specific working implementation. That distinction makes his achievement more interesting: a published possibility still needed someone to turn it into something usable.

First, get it into the world

The Ethereum Foundation’s August 2018 grants announcement listed $100,000 for Uniswap, describing it as a decentralized-exchange framework. That support bought room to work on a production release. Security review mattered. Software that handles assets needs more scrutiny than software whose worst failure is an unfortunate button color.

Runtime Verification worked on formalizing and checking the contracts. Callil Capuozzo, Adams’s friend since school, designed the interface. Kyokan engineer Jacky Chan built the launch interface. Those contributions complicate the familiar picture of a solitary programmer creating a market. The program had an author; getting people to use it required several kinds of expertise.

The shirts were part of the launch, too. Capuozzo and his brother Leander made them by silk screening; Capuozzo carried the finished batch to Prague. A public protocol could be abstract. A shirt was something a conference attendee could wear home.

At launch, one provider put roughly $30,000 into pools across three tokens, enough for swaps of about $100. This was a modest beginning for a financial venue. It was also an actual beginning: contracts deployed, an interface available, and assets there to exchange. Adams’s tweet went to roughly 200 followers. The thing he had learned to build was now available for other people to judge.

Success arrived with company

Uniswap v2 reached Ethereum in May 2020. Among its changes were direct pairs between ERC-20 tokens, rather than requiring each pool to pair a token with ETH. It also added price oracles and flash swaps. These changes expanded what developers could do with the protocol and gave traders more useful paths through it.

By the end of August, the scale had changed. Uniswap recorded $426 million in daily trading volume against Coinbase’s $348 million in a widely reported comparison. One day’s volume is a snapshot, not a permanent ranking. For a project that had begun as a programming exercise, it was still a remarkable change in the size of the comparison.

Competition followed. SushiSwap used Uniswap’s publicly available code and offered its own token incentives. A competitor could borrow the machinery and try to attract the capital. Adams’s work was becoming a practical test of openness: the same conditions that let other developers build around Uniswap also let them compete with it.

“Competition has made us work faster.”

Hayden Adams, September 2020

On September 16, 2020, Uniswap introduced UNI. Eligible historical user addresses could claim 400 tokens each, including addresses whose only interactions had been failed transactions. The announcement’s affectionate acknowledgment of that last group was a small joke with a material consequence. Even an unsuccessful attempt to use the exchange could count as participation.

UNI brought governance into the story. A protocol can execute a swap through predefined rules; deciding how an ecosystem should develop requires debate, proposals, and votes. Adams now had to build in a setting where users, investors, liquidity providers, and independent developers could have different priorities.

More choices, more responsibility

Uniswap v3 introduced concentrated liquidity in 2021. Providers could choose the price ranges in which to put their capital to work. Instead of distributing liquidity across every possible price, they could focus it where they expected trading to happen. The design made capital more useful within those ranges, while asking providers to make more deliberate choices.

That tradeoff is a recurring feature of Adams’s work. A simpler experience for the trader can rest on more sophisticated decisions elsewhere. The interface may offer a short route from one asset to another. Underneath sit choices about liquidity, pricing, transaction costs, and the people who supply the assets.

In October 2022, Adams announced a $165 million Series B led by Polychain Capital. He described a broader product effort encompassing the web app, developer tools, and mobile access. An engineering experiment had become a company with capital and customers. His stated aim remained making the protocol’s capabilities accessible to more people.

With v4, launched on January 31, 2025, developers gained hooks: modules for customizing pools, swaps, fees, and liquidity positions. Nine independent audits preceded the release. Unichain followed on February 11, an Ethereum layer 2 designed for financial applications. Adams’s team was now working on both the exchange and the infrastructure around its transactions.

Publishing code, answering questions

Visibility also put Adams in public arguments about regulation. After Uniswap Labs received an SEC Wells notice in April 2024, he appeared on Bankless to defend the company’s position. A central question was whether publishing software that runs on a decentralized blockchain should amount to operating an exchange. Adams argued that it should not.

In February 2025, the SEC closed its investigation without taking action. The episode added a different kind of work to the founder’s job: explaining the relationship between a software company, the interface it builds, and a protocol that other people can access independently. Those distinctions are easy to compress into a brand name. They matter a great deal when responsibility is being assigned.

The next market may look familiar

By August 2026, Adams was making a broader argument about tokenization. His essay on correlated pairs asked whether automated market makers could become an engine for financial markets beyond crypto. Tokenized assets, he argued, could trade directly against other assets whose prices tend to move with them, rather than routing everything through dollars.

His example was a tokenized Nvidia share paired with a tokenized S&P 500 fund. The attraction, in his account, is that an investor who already wants to own both assets may supply liquidity differently from a professional firm that must hedge its inventory. This is Adams’s thesis about how markets could develop. Its appeal depends on whether the economics work for the people committing capital.

There are concrete projects to watch. In October 2026, SMBC Nikko and Nethermind announced work on a Japanese DeFi Gateway using Uniswap v4, with Labs supporting development. Completion was targeted for mid-2027. Later that month, Aggregator Hooks went live, bringing external liquidity sources into routes accessible through Uniswap’s web app and API.

These developments give Adams a larger version of his original problem. An idea about markets needs usable software, willing participants, and enough patience to find out where it fails. The engineer who once needed a friend to suggest a project now leads a company building several of them. There will still be announcements to write. Fortunately, those remain editable.