The Unbundling Of Ethereum

Software is the encoding of human thought, and as such software is limited only by our collective imagination. Ethereum embraced this as the first computer that allowed anyone to program arbitrary logic on a trust-minimized, global, permissionless, censorship-resistant computer.

This unlocked an incredible wave of experimentation. How could anyone have forecast the massive opportunities for Open Finance and Web3 in 2014?

Traditionally, most software applications lived on separate databases. It was actually counterproductive to bundle most types of software for technical reasons. However, in the case of a trust-minimized world computer, the opposite appears to be true. By design, Ethereum bundles every app on a single, global state. By bundling such a wide variety of applications— from gambling games to cats to open finance and more—Ethereum has earned a monetary premium, defined as the excess wealth stored in ETH, above and beyond what’s necessary for simple gas payments.

As the utility of ETH has increased with a growing number of applications built on Ethereum, it’s possible to quantify some of the key enablers that give ETH a monetary premium.

1. The security budget for the Ethereum ecosystem as a whole grows.
2. Liquidity increases, making it easier for market participants to utilize ETH.
3. The power of the single schelling point that is the Ethereum blockchain grows (e.g. brand value, integrations with 3rd parties).

Today, Ethereum is the leading smart contract platform. This is primarily because it was the first of its kind. However, although Ethereum is theoretically capable of encoding any conceivable logic, in practice it’s rather limited because of throughput, latency, and cost constraints. While the trust-minimization properties of Ethereum are compelling, many developers are realizing that Ethereum 1.x, as a proof-of-work (POW) chain running the Ethereum Virtual Machine (EVM), simply cannot support the applications they want to build.

Because of these limits, developers are choosing to build on other chains at an increasing pace. It’s indisputable at this point that the Web3 stack is becoming more heterogeneous across many chains rather than homogenous on top of just Ethereum.

What applications are being unbundled from Ethereum, what applications are increasingly bundling around Ethereum, and what are the long-term implications for ETH’s monetary premium?

Let’s begin.

Unbundling By Vertical

In its earliest days, a large percentage of Bitcoin traffic was gambling, e.g. Satoshi Dice. Some of that moved to Ethereum (e.g. FunFair, EtheRoll, DAO.Casino, Edgeless Casino, iDice), but almost all of the gambling apps have migrated to Tron. The gambling use case does not require the trust-minimization guarantees that Ethereum offers, and so blockchains like Tron actually offer a better platform for these applications than Ethereum because they trade trust-minimization (DPoS vs PoW) for UX (e.g. transaction confirmation time).

Meanwhile, game developers are building an array of games that leverage the blockchain for in-game asset ownership. Again, this use case doesn’t require the trust-minimization guarantees that Ethereum provides. That is, while Ethereum’s validators offer strong guarantees with regards to the validity of state updates, the vast majority of in-game items are worth less than $100. Even if block producers in Tron violate user trust and steal those items, the maximum downside on a per user basis is… $100. Gaming is taking off right now on many non-Ethereum chains including Tron, NEAR, and others, and a number of high-profile venture-backed studios have launched that are building blockchain-based games on other platforms including ITAM games, Mythical games, and many more.

Perhaps the most successful early use case on Ethereum was capital formation for open source software projects in the form of initial coin offerings (ICOs). On a technical basis, this requires the creation and issuance of tokens, and not much else. Binance has made strong inroads on this front in 2019 with their Launchpad platform. And with the launch of Binance Chain, Binance is making an aggressive push to become the best platform to support global capital formation for open source software development. Binance Chain is designed from the ground up to support digital asset issuance, payments, and trading. Although Binance Chain is still early, it’s clear that Binance Chain aims to unbundle one of the key uses for Ethereum.

Given the success of ICOs on Ethereum, it was only natural to expect more traditional forms of capital formation—in the form of registered securities—to be issued on the Ethereum blockchain by using Ethereum’s programmability to support transfer restrictions necessary to support regulated securities. However, this use case appears not to have found product/market fit, as Ethereum-based securities exchanges have quietly shut down, and securitized real estate offerings have failed. While at least one security token issuance is migrating to Tezos, the problem seems to be lack of product/market fit. Because securities are not bearer assets, they simply don’t require the trust-minimization properties that blockchains offer. Meanwhile, Carta, a centralized repository of private equity, continues to grow unabated.

After making some noise about the Enterprise Ethereum Alliance, traditional financial institutions are moving away from Ethereum. JP Morgan is aggressively pushing its own permissioned JPM chain, and alternative investment managers like USCF are moving to Kadena because of the safety benefits of its Pact programming language. Other fintech Fortune 500s will launch on Kadena later this year as well.

Given Ethereum’s success in issuing and managing digital assets, it’s only natural to assume that Ethereum will be used to manage digital assets for real-world events. However, the Tari team is building a full-stack digital asset platform for real world events as a separate platform. They’re starting off by focusing on tickets for concerts and live events. The Tari team built Big Neon, a consumer-facing application that’s live in three cities now, and will be live in many more by the end of the year.

Terra just launched its blockchain using the Cosmos SDK. Given the team’s background and connections in Korean e-commerce, and the Korean populace’s general affinity for crypto (Koreans account for 30% of global crypto trading), it’s likely that the Terra blockchain will become the most widely used blockchain for retail commerce. If this gains traction, they are likely to expand into brick-and-mortar commerce, and into neighboring countries. As they grow from the base of Korean e-Commerce, they are likely to bundle more into their own chain.

Helium tried to build on Ethereum in 2017, but realized that Ethereum simply could not serve their needs. Instead, Helium built their decentralized network for local LP-WAN nodes using their own highly optimized blockchain. This blockchain is rolling out in select cities soon.

Kik Messenger launched their Kin ICO on Ethereum, and quickly realized that Ethereum would not support the sheer volume of users on their platform. Today more than 250,000 are using Kin across more than 40 applications in the Kik ecosystem.

TOP Network tried to build on Ethereum, and found it unsuitable for their applications.

Oxio, which is building a marketplace for telecom data exchange for mobile virtual network operators (MVNOs), considered Ethereum, but ultimately chose to build on Stellar because Stellar offers fast, inexpensive transactions and a native DEX.

I’m sure there are many more examples that I’ve missed, but I think the point is clear. Ethereum simply cannot serve the vast array of trust-minimized applications at global scale. Developers are realizing this, and are going elsewhere.

Open Finance

However, this is not to suggest all doom and gloom. Ethereum appears to have at least one use case that’s working that no other ecosystem has in a meaningful capacity: open finance (AKA DeFi). It’s worth considering why:

First, in order for open finance to work, it’s intuitive that the aggregate market value of the layer 1 currency needs to be sufficiently large. Or stated another way, it’s not really possible to support an open finance ecosystem when the aggregate economy is worth less than $1B. None of the other smart contract platforms (yet) store enough wealth and liquidity to meaningfully compete with Ethereum on this front. This logic is circular: Because Ethereum is one of the only chains that meets this threshold, and because this use case is presumably value accretive to ETH’s monetary premium, this creates a unique moat for Ethereum.

Second, open finance can work in spite of Ethereum’s shortcomings (high transaction fees and high latency). For example, users will happily wait one minute and pay $1 or $2 in gas fees to take out a $25,000 loan on Maker or Compound. This experience is still 100x better than applying for a loan through the legacy financial system.

Given these realities, it’s natural that the Ethereum ecosystem has coalesced around open finance.

Ethereum 2.0

By separating applications into separate shards, the system as a whole will gain throughput and reduce transaction fees. However, given the permissionless nature of sharding, it’s not actually clear how applications will function themselves across shards.

For example, does the 0x contract live on all shards (and pay state rent on all shards), or just some shards? Given the perpetual risk of global settlement, will the Maker contract allow users to move DAI from shard to shard? If there are 10M daily prediction market users and a single shard can only accommodate 1M users, how will the Augur contract manage markets across shards and deal with outcome reporting? How will ENS work across shards?

While these are likely solvable problems, they aren’t going to be solved anytime soon. These problems are intrinsically second- and third-order problems that arise from sharding, and the developer community at large is only going to be able to understand all the nuances and experiment with solutions once they can actually deploy applications in a sharded blockchain environment. Once deployed, developers will learn about user and data flows across shards, and then develop the right models and abstractions to accommodate real world use cases.

Ethereum launched in Summer 2015. The first open finance application that went live was probably EtherDelta in July 2016, followed by 0x and Maker in late 2017. It took about 2 years to build the first unsharded DeFi applications on Ethereum 1.0. It’s hard to expect this cycle to proceed much faster in Ethereum 2.0 given the additional complexity that sharding introduces while maintaining modularity and composability (as I argued in the Crypto Mega Theses, modularity and composability are the primary enablers of Open Finance).

Note that these challenges are not unique to Ethereum 2.0. Any sharded application environment, regardless of underlying security model and latency assumptions, faces these logical problems. Sharding is just really complicated, and while it may be the right scaling solution on a long enough time scale, it introduces incredible new complexity.

Building New Bundles

For the next few years, we should expect the pace of Ethereum’s unbundling to accelerate, especially in the absence of a usable Ethereum 2.0. Until the standards that developers use to build trust-minimized applications are widely accepted and understood across the full range of sharding and layer 2 solutions (side chains, state channels, zk-verified off-chain computation, etc.), the market will become increasingly heterogeneous rather than homogenous.

ETH has developed a significant monetary premium because it bundles apps, liquidity, security, and provides a global schelling point. While it’s unlikely that open finance moves away from Ethereum anytime soon, it will be interesting to see if ETH retains its monetary premium as other smart contract platforms gain adoption. On one hand, there’s no reason to think that these new chains will adversely impact ETH in the short run. On the other hand, if they present a credible long term threat, the market should rationally discount ETH’s monetary premium from what it is today.

I suspect that because most of the new chains are unlikely to challenge Ethereum in open finance - which is intrinsically a money-centric application - the next generation chains are unlikely to impact ETH’s monetary premium in the foreseeable future. Each of the new platforms, in order to bootstrap themselves, is going to try to pioneer new markets rather than compete directly with Ethereum where Ethereum is strong today. Thus, I expect that that an increasing number of smart contract platforms will develop some monetary premia over the next couple of years rather than detract from ETH’s.

For now, both BTC and ETH have been able to maintain their respective monetary premia because the underlying platforms are so differentiated. However, as other smart contract platforms grow, they will bundle applications and develop their own respective monetary premia.

Credit to Multicoin’s analysts, Spencer Applebaum, Ryan Gentry, and Ben Sparango for the conversation that sparked this essay, to Chris Dixon for some of the ideas and analogies used, and to Albert Wenger for providing feedback on this post.