Unit Economics of Decentralised Exchanges
Exchanges mostly facilitate decentralised Finance (DeFi). Within the realm of exchanges, there exists a spectrum of (de)centralisation. On one end, we have the automated market makers (AMMs) pioneered by Hayden Adams, the founder of Uniswap, which embody decentralisation. On the other end, we have order books, which resemble the structure of traditional exchanges but are not as decentralised.
The overarching goal of these exchanges is to eliminate intermediaries and enable users to trade directly from their wallets in a non-custodial manner. Order books operate based on a trade execution model that matches buy orders (bids) with sell orders (asks). In contrast, the AMM model utilises pools of crypto tokens supplied by liquidity providers (LPs) to determine prices through an algorithm that adjusts based on the changing token ratio within the pool.
This research report aims to compare two major types of DEX models: order books and AMMs. The focus of the analysis will be on their respective volumes, fees, trader unit economics, and earnings. The data used for this analysis is derived from the top five DEXs of each type. However, it is important to note that the figures may be influenced by prominent platforms such as Uniswap and dYdX.
When examining the volumes of order books and AMMs on an annualised basis, it becomes apparent that order books have experienced a relatively minor decline and stagnant growth over the past two years. Conversely, AMMs have encountered an average decline of 34% during the same period. It is worth noting that Uniswap, the leading AMM platform, and dYdX, the leading order book platform, contribute to approximately 75% and 90% of their respective categories’ total volumes. This significant presence of these platforms has a substantial impact on the overall volume trends. It is interesting to highlight that dYdX is one of the few exchanges with a year-over-year volume increase.
However, when it comes to fees, the story diverges. Order books account for approximately 65% of AMM volumes but contribute only around 15% of the total fees generated. This discrepancy can be attributed to the fact that fees are considerably cheaper on Layer 2 solutions, where dYdX primarily operates, which bestows a significant portion of order book fees. On the other hand, AMMs predominantly operate on Layer 1 blockchains such as Ethereum, Binance, and Avalanche, with some utilisation of Ethereum Layer 2 solutions. These Layer 1 blockchains tend to have less scalability, higher dollar-denominated fees, higher volumes, and deeper liquidity, leading to increased fees.
We project order books to achieve approximately $6 million in profits by the end of this year, whereas AMMs collectively face a loss of $330 million. The trend observed here is that revenues, akin to operating profits in traditional financial markets, decreased at a lower rate for order books, while token incentives, similar to operating costs, decreased at a higher rate than AMMs. The business model of AMMs, where a significant portion of fees is directed to LPs rather than the protocol itself, makes it unlikely for them to become profitable with their current architecture and infrastructure. This is compounded by the fact that communities are worried about US regulations, such as in the case of Uniswap’s community mulling the fee switch proposal. Theoretically, Uniswap could have made around $19m in profit (see our investment case here), pushing the whole AMM category in the green. However, that is not the case.
The unit economics of order books and AMMs provide valuable insights into profitability and the value generated per user. From a net profit perspective and per-trader basis, order books emerge as the more profitable business model, showcasing higher lifetime value and earnings per user. This indicates that order books have been successful in generating more fees per user and maintaining better user retention rates compared to AMMs.
In summary, AMMs exhibit greater decentralisation and benefit from deeper liquidity and higher volumes, partly driven by their first-mover advantage and the trust traders place in battle-tested, safer, slower Layer 1 blockchains. While these DEXs command high fees from users, their inefficient business models, where LPs require significant compensation, result in minimal redirection of fees to the protocol, leading to sustained unprofitability. On the other hand, order book models prioritise speed and cost-effectiveness, which are more challenging to achieve on-chain. Nevertheless, the efficiency of these models makes the unit economics per trader more attractive. This year’s emerging profitability trends are further bolstered by the fact that order book volumes are catching up to AMMs, with notable contributions from platforms like dYdX.
