Author: Mason Nystrom, Unsyndicated, former Variant Fund investor; Translated by: 0xjs@Golden Finance

Over the past 30 days, Ethereum’s order flow has exceeded $25 billion, with nearly half of that order flow coming from proprietary applications.

As the value of blockspace commodities continues to grow and pave the way for fat applications, the privatization of order flow will only grow.

Source: Orderflow.art

But how did we get here? Where are we going?

The short answer is how we got here, food tokens. Long story short, Defi Summer gave rise to a large number of prosumers and retail trading, which subsequently gave rise to trading aggregators such as 1inch, which provided users with better price execution through private order routing. Wallets (such as MetaMask) soon followed suit, realizing that they could monetize user convenience by adding in-app exchanges, proving that any application that controls the attention (and orders) of end users has a very valuable business model.

Source: Dune

Over the past two years, we have seen two additional categories of players enter the private order flow space - Telegram bots and solver networks. Telegram bots align with MetaMask's "convenience fee" and provide users with an easy way to easily trade long-tail (shitcoin) assets in group chats. As of June, Telegram bots account for approximately 21% of private order flow transactions and 11% of volume, most of which are conducted through private mempools.

On the other hand, on the market’s big end, solver networks (i.e. Cowswap, UniswapX) have also become the core venue for trading a large number of highly liquid currency pairs (such as stablecoins and ETH/BTC). Solver networks have changed the order flow market structure by outsourcing the work of finding the best path for a given trade to a competitive market of solvers (market makers).

As a result, trading venues have bifurcated, with convenient frontends (including TG Bots, wallet exchanges, and Uniswap frontends) primarily used for long-tail, low-value (less than $100k) trades, while aggregators and solver networks are the preferred venues for larger trades, typically involving stablecoins and major currencies (ETH/BTC).

With more granular filtering, you’ll notice that most of the private order flow comes from the frontend (TG bots, wallets, and frontends).

The privatization of order flow is even more pronounced when we consider that on a trade basis only 15-30% of Ethereum transactions go through private mempools, meaning that a small number of exchanges contribute a large portion of private order flow.

Source: Dune

In other words, valuable order flow is more important than the amount of order flow. The power law of users and order flow leads to an inevitable conclusion - applications will account for the largest proportion of total value. In other words, the fat app theory is still valid.

Towards fat applications

Uniswap’s protocol is obviously valuable, but the more interesting story is happening at the application layer, as Uniswap strives to become a consumer app — a key component of verticalizing its stack — starting with expanding the functionality of its interface, mobile wallet, and aggregation layers. For example, Uniswap Labs’ app — Uniswap’s frontend, wallet, and aggregator, UniswapX — accounts for nearly 25% of the $13 billion in private order flow over the past 30 days, and nearly 40% of total order flow (private and public).

Elsewhere in the cryptocurrency space, applications such as Worldcoin account for nearly 50% of Optimism mainnet activity, which has pushed them towards launching their own app chains, further emphasizing the power of the fat app theory and control requirements (such as users and transactions).

Even top NFT projects with strong brands like Pudgy Penguins are building their own chains, with Pengu CEO and CEO Luca explaining that controlling the distributed block space is conducive to the value accumulation of Pudgy's brand and IP.

Going forward, applications should look to create new types of order flow, whether by creating new assets (e.g. Pump and memecoins), by building applications that create new user utility (e.g. identity) (i.e. Worldcoin, ENS), or by crafting better vertically integrated consumer experiences and supporting valuable transactions like Farcaster and frames, Solana Blinks, Telegram and TG applications, or on-chain games.

Final Thoughts on Fat Apps

It is worth noting that the fat application theory has become the focus of many people in the cryptocurrency field. Since the end of the last cycle, the application chain theory has become part of the consensus view.

My current take on the fat app theory is that we will see the majority of value accrue to the application layer of the stack, where control over users and order flow puts applications in a privileged position. These applications will likely be coupled with on-chain protocols and primitives, similar to today’s UniswapX and Uniswap Protocol, Warpcast and Farcaster, Worldcoin and Worldchain. Ultimately, these protocols, especially those that are most on-chain (e.g. MakerDAO), can still accrue significant value, but applications will likely capture more value given their proximity to users and off-chain components that provide more defensible moats for applications.

In the end, I still believe there is a path for Layer 1 blockchains (e.g. Bitcoin, Ethereum, Solana) to gain significant value as non-sovereign reserve assets, while the underlying assets (e.g. ETH) accrue significant value. Given enough time, applications may try to build their own L1s, just as they have built their own L2s, but building L2 blockspace commodities is very different from bootstrapping L1s and turning tokens into commodities and collateral assets, so this may be a distant future.

The core takeaway is that as more and more consumer applications create and own valuable order flow, the crypto world will re-evaluate applications as people come to the inevitable conclusion - fat apps are inevitable.