A Guide to Understanding ETH as an Investment

Celia Wan
Dragonfly Research
Published in
15 min readJun 29, 2021


By Kevin Hu and Celia Wan

This piece is an attempt to help readers understand the potential impact of the Ethereum network and its native asset, ETH, as an investment. It does not try to arrive at a precise valuation or price prediction. It assumes the reader has a basic understanding of Bitcoin, Ethereum, and the broader crypto ecosystem. The views expressed in this document are not, and should not be construed as investment advice.


Launched in 2009, Bitcoin became the first trustless, digital-native money. Due to its scarcity and unforgeability, it has the potential to be a universal non-sovereign store-of-value. With the flood of institutional adoption in the second half of 2020, Bitcoin’s narrative as digital gold solidified. Eleven years after the launch of Bitcoin, institutional investors are finally investing in Bitcoin as this simple yet powerful Store-of-Value narrative begins to go mainstream.

On the other hand, investors’ understanding of the Ethereum network and its native asset, ETH, is convoluted. Smart contracts, Decentralized Finance, and Web 3.0 are still niche topics. ETH as an asset has a complex narrative that even people working full-time in crypto have a hard time articulating.

By design, ETH is the gas for the Ethereum network, which is popularly described as the “world computer”. But what is the “world computer” good for? Is ETH a consumable commodity, capital asset, or programmable collateral/money? How does ETH accrue value?

A singular ETH narrative has yet to crystallize. The rest of the essay will explore the different emerging perspectives around ETH and hopefully help investors understand what they are potentially betting on when they invest in ETH.

A Popular (But Slightly Dated) Economic Analysis of ETH

The current version of Ethereum is akin to a distributed operating system with a native token, ETH, that pays for the cost of computing. The miners who provide the computation resources are rewarded through the block rewards and transaction fees. Historically, ETH block rewards have been adjusted with the explicit goal of targeting a particular spend, giving ETH a much more activist monetary policy compared to Bitcoin.

In this current model, Ethereum users pay the transaction cost in ETH, and ETH holders bear the cost of inflation. Implicitly, in absence of speculation, ETH holders are betting that demand for ETH from the users of Ethereum-based applications outpaces the rate of inflation through block rewards (which has fluctuated wildly in the history of Ethereum).

A popular economic analysis of ETH is to view the Ethereum network as an economy and thus express the network’s total value as GDP. We can then calculate Ethereum’s network value based on the equation of exchange PQ = MV, whereby the price times quantity (total output) must be equal to the money supply multiplied by the velocity (turnover) of money. Therefore, the total value of the Ethereum network is equal to the total amount of ETH in circulation multiplied by the turnover of ETH.

A common conjecture is that, in the long-run, the GDP of the Ethereum network is likely going to be reasonably sizeable but not massive, as it is limited by the deflationary nature of computing costs and the fact that Ethereum must be very cheap to use to gain mass adoption. The turnover of ETH will also likely be extremely high because there is no reason to hold such a frictionless means of payments. As a result, the valuation of ETH can be relatively low and still be able to support a reasonably sized economy. In this framework, ETH is unable to capture significant economic value because it is a perfectly replaceable commodity that does not need to be held.

This argument is centered on 1) that ETH’s only function is as a means of payment; 2) the open-source nature of public blockchains make it impossible for them to retain IP value; and 3) the cost of switching blockchain is near zero for applications. The result is that Ethereum will have a weak network effect and ETH should be priced like a commodity as users will not be willing to pay more than the cost of production.

Recent Developments

While the logic behind the PQ = MV argument is sound, the thesis has not played out so far. Ethereum is still by far in the lead in terms of usage and developer adoption. Its network effect does not seem to be weakening and the market cap of ETH is 5x larger than the third most valuable Layer 1.

Ethereum today is vastly different compared to two years ago. In 2017 and 2018, the only use case of the Ethereum blockchain was to issue ICOs for projects (the vast majority of them have not delivered anything meaningful). Today, Ethereum supports a flourishing Decentralized Finance (DeFi) ecosystem, and other use cases such as NFT/gaming/metaverse and Web 3.0 are beginning to develop.

Ethereum average gas prices (as proxy for usage; Source: Etherscan)

DeFi became the first widely used application of the Ethereum network. Currently, DeFi has $60B in AUM (peaked at $120B+ in early May 2021), $17B+ loan outstanding, and facilitates, on average, $5B in trading volumes a day. Together, these DeFi applications generate $4.5B+ annual revenue (last 30 days fees paid to protocols as a proxy for revenue) and have built up substantial liquidity to create network effects.

The current DeFi ecosystem on Ethereum has formed a positive feedback loop, where users both bring the liquidity-based network effect to DeFi and benefit from a liquid DeFi market. As more assets are locked in DeFi, the slippage on Automated Market maker (AMM)-based decentralized exchanges and the cost of borrowing on lending platforms decrease, making it more attractive for users (though there is no network effect across pools, and aggregate network effect is sublinear vs dollars locked).

Furthermore, DeFi protocols’ composability and interoperability with each other have created a lock-in effect for Ethereum, making it hard for other Layer 1’s and side-chains to compete. To compete with Ethereum, other chains need to foster an entire DApp ecosystem and source liquidity from scratch, which is not an insignificant undertaking and usually requires heavy token subsidies. In fact, we have seen chains such as Polygon and BSC achieve success via these routes by replicating DeFi applications on Ethereum while offering additional appeal of low gas fees and liquidity mining programs. Even so, however, bridging these different blockchains still breaks composability between DeFi protocols and therefore forces Layer 1’s to develop their separate ecosystems.

Total Value Locked in DeFi (Source: DeBank)

From an architecture perspective, the most significant upcoming changes for Ethereum are Ethereum 2.0, Layer 2’s, and EIP-1559. Without going into technical details, Ethereum 2.0 will transition the Ethereum network from a proof-of-work consensus model to a proof-of-stake consensus model. Instead of using computational resources to verify the next block, Ethereum 2.0 will rely on ETH holders (validators) to vote on the next block, making the security model based on game theory and economics. Ethereum 2.0 will also split the network into 64 shards that will parallelize the network and allow for it to scale in the future.

In the near term, Layer 2's, specifically rollup-based approaches based on fraud proof and zero knowledge proof, may make the Ethereum network more scalable and cheaper by orders of magnitude. Ethereum today handles about 15–17 transactions per second (tps). We expect the upcoming launches will increase Ethereum’s throughput up to 2,000+ tps (with on-chain data) and 9,000+ tps (with off-chain data) on a Layer 2. Layer 2’s will also lower transaction costs significantly. A typical DeFi interaction today costs $10 — $100+, which crowds out the most applications and users. Layer 2’s will likely reduce the cost of a DeFi transaction to under $1 (some estimates are as low as 10–20c), making DeFi and Ethereum 100x more accessible.

Lastly, Ethereum is likely to adopt a new monetary policy with an update called EIP-1559. EIP-1559 introduces several changes to Ethereum’s transaction fee algorithm. For the purpose of this analysis, the most important upgrade would be that EIP-1559 results in most of the Ethereum transaction fees being burned instead of paid to miners. If proof-of-stake and EIP-1559 are implemented, ETH’s inflation will likely decrease meaningfully and ETH will become a capital asset in addition to a consumable commodity. This should have a profound effect on ETH’s value accretion. More on this later.

Where Are We Now?

As impressive as the story we have told so far about Ethereum, it is worth noting that the traction DeFi has generated so far lives rather precariously on a limited group of users, taking advantage of ephemeral liquidity mining programs and project IDOs. However, the various developments on Ethereum does make the PQ = MV model of evaluating ETH’s value obsolete, and new ways to conceptualize ETH are needed to accommodate the rapid development on the Ethereum network.

The following three sections introduce different ways to conceptualize ETH both qualitatively and quantitatively. These conceptualizations will then together contribute to the final evaluation of ETH. It is worth reiterating that neither the mechanisms nor the final evaluation described below should be taken as the final verdict on ETH. Nevertheless, it offers an honest attempt to understand ETH as an asset and Ethereum as a platform.

Ethereum as the Financial Layer of the Future

Ethereum made contracts auto-enforceable without trusted third parties. Its various token standards allow value, other than ETH, to be represented on the network. Together, Ethereum enabled value and ownership to be governed by code, which can serve as an alternative to the existing trading and settlement rails of traditional finance. As the technology improves and transaction cost declines, Ethereum and DeFi will enable new use cases that are unimaginable today.

We believe Ethereum and DeFi have the potential to disrupt legacy finance and should be valued accordingly because of the following:

  • Permissionless innovation at software speed: all DeFi protocols are open-sourced and composable. DeFi entrepreneurs can remix and innovate in finance at incredible speeds and reach a global audience with almost no fixed cost.
  • Incentive alignment: with the right token mechanism design, all the stakeholders in the ecosystem (protocols, users, liquidity providers, engineers/protocol maintainers) can be properly aligned and growth can be bootstrapped with almost no upfront cost.
  • Cost reduction: DeFi eliminates costs associated with legal, labor, compliance, and fixed infrastructure. Financial transactions in the legacy world are based on legal constructs and enforced by the government. The cost of recourse can be high and non-transparent in complicated situations like bankruptcies. These costs do not exist in DeFi because value is governed entirely by code.
  • Frictionless capital and near-instant settlement: Capital within DeFi flows frictionlessly and programmatically with the sub-minute settlement — it is a truly digital-native experience. In comparison, payment rails from different jurisdictions do not natively talk to each other (better now with companies like Stripe and Plaid) in the legacy world. The processes are manual and many systems are decades old.
  • Mass customization & synthetic assets: much like how the internet enabled companies to reach different niches users, DeFi will enable users to access almost any asset in the world. Today, through AMMs like Uniswap, anyone can create a new trading pair as long as they have the trading assets in inventory. Synthetic assets are taking this to another step. Theoretically, as long as there is a trusted data source, anyone can create a synthetic asset by providing on-chain collateral. Teams at Synthetix, UMA, Mirror and many others are exploring in this direction.
  • Governmental neutrality: the financial system built on Ethereum is open and accessible to all. While this may not sound very attractive to users in countries with matured financial markets such as the U.S., it offers unparalleled advantages to people who live in countries where the local financial systems are less efficient or corrupted.

ETH as Capital Asset

ETH’s utility as the default form of payments for the Ethereum network activities sanctions a rather straightforward value-based pricing model for estimating the asset’s valuation. This is especially true once the network scales with Ethereum 2.0 and EIP-1559. Even if other tokens such as stablecoins may replace Ethereum in the future as alternative forms of payments, ETH will still be used by network validators for staking and receiving rewards. Therefore, ETH will likely accrue value from demand like a consumable commodity and cash flows like a capital asset.

Let’s say, in 10 years, Ethereum facilitates $5 trillion of transaction volume across 4 billion transactions per day. $5 trillion per day is underwriting a 74% annual growth rate. This seems reasonable if you believe Ethereum is going to grow like the internet in its early days.

The number is even arguably an underestimation if you consider Ethereum’s growth track record. Ethereum is currently handling $8B of ETH and $10B of stablecoin volumes per day (this is an underestimate of total volume as we excluded other ERC-20 transactions), which are a 6x and 15x increase respectively from 2019 to 2020.

In this scenario, transaction count is the more aggressive assumption (1.2M tx to 4B tx per day is 125% annual growth rate), but could be reasonable if Ethereum is cheap and scalable enough to have programmable micropayments between contracts. Ethereum, after more than 3 years, is finally at the cusp of launching scalability solutions. As we highlighted before, the cost of DeFi transactions could decrease by 20–50x in the next 6 -12 months, which could increase its transaction count exponentially.

It is also conceivable that there will be a distinction between high-value transactions (not necessarily in dollar value transferred, but potentially in extractable value), which should generate fees based on value transacted (fees are % of the transaction) and low-value transactions, which should be commodity-based (fees vary slightly depending on network utilization).

High-value transactions could potentially be value-based because ordering within an Ethereum block matters, especially if the transaction allows the user to make an economic profit. Today, high-value transactions are often front-run, which are near impossible to prevent (see Ethereum is a dark forest) unless the transactions are sent to a private mining pool (who will likely charge value-based pricing). Several teams are working on solutions to build an orderly marketplace for extractable value, which could be implemented in the medium-term.

However, it is likely that there will always be transactions (such as arbitrages, which is highly competitive) whose value will far exceed the base fee and therefore make paying excess tips for prioritized finality lucrative. Although these transactions would only account for a very small portion of the total network transactions.

In EIP-1559, the value-based fee is formalized in the form of a tip to miners/validators (may potentially be not ETH denominated), whereas the commodity-based fee will be the base fee that is denominated in ETH and gets burned. Unlike in Ethereum 1.0 where fees go to miners, both the base fee and tip will accrue to ETH holders in EIP-1559 if Ethereum transitions to a proof-of-stake consensus model. The base fee creates deflation and generates demand for ETH while the tip (regardless of denomination) is like cash flow to miners/stakers who are ETH holders.

Daily Tx Fee (7D avg) on Ethereum (Source: The Block)
Daily ETH Transfer Volumes (7D Avg; Source: The Block)
Chart ETH Tx Count Incl. ERC20 & 721 (7D Avg; Source: The Block)

Let’s do some back of the envelope math using the hypothetical numbers above and assume the following (these numbers are purely for illustrative purposes):

  • 5% of the $5 trillion will be high-value transactions priced at 0.05% of transaction value. The majority of this will likely be from DeFi.
  • The base fee averages out to be $0.01 per transaction on the 4B tx per day
  • At this stage, Ethereum will be fairly mature and fee growth will diminish to 5% a year in perpetuity.
  • The economy’s long-run risk-free rate in the distant future is 2%. ETH stakers will require another 5% to compensate for the risk of slashing and technical risk (these risks will have stabilized in the long-run hence 5% is sufficient), making the long-run required yield of ETH 7%.

Using the assumptions above, Ethereum will generate $60.2B annual fees. This is only an 18x increase from today whereas transaction volume and count would have increased by 250x and 3300x respectively. The deflationary power of technology!

Finally, pricing ETH like a capital asset using a dividend growth model with the $60.2B number above gives the token a terminal value of $3.2 trillion in 10 years. Although this may seem like an enormous number, it is ten years out and does not account for any risk (e.g. technical, competition, regulatory). A shrewd investor needs to apply the appropriate probability or discount rate.

ETH as Monetary Asset

ETH derives its value from two sources. First, its utility value, as described in the argument above. But the second source of its value is in its monetary premium, which derives in its preference as a “money-like” asset for use in the Ethereum economy.

As a medium of exchange and unit of account, ETH is unlikely to succeed. An ultra ETH bull may make the case that if the Ethereum network becomes ubiquitous and fees stabilize, ETH can become a popular currency. In general, that is highly unlikely and stablecoins are fundamentally much better at both functions even if Ethereum does become a dominant platform. Today, the total on-chain volume of ERC-20 stablecoins already exceeds that of ETH ($10B vs $8B per day) despite ETH having 5x the market capitalization. The difference between stablecoins and ETH off-chain volumes are even larger.

As a non-sovereign store-of-value, however, ETH may still have a chance at taking some potential market share through being used as collateral in DeFi.

In the long-run, it is conceivable that ETH can even compete with Bitcoin on the dimensions of scarcity, durability, and unforgeability for the following reasons:

  1. ETH’s monetary policy will stabilize and inflation will likely be cut in half (from 4% to 2%, according to Tim Roughgarden) because of EIP-1559. Of course, this is not the same as having a fixed supply, but a well-telegraphed low inflationary asset may be the next best thing.
  2. The security model of Ethereum 2.0 will eventually be almost just as battle-tested as Bitcoin’s (in 20 years, Bitcoin will only have existed 20% longer than Ethereum); Additionally, Ethereum 2.0 and proof-of-stake will likely increase the security guarantees of Ethereum if ETH is sufficiently valuable (we acknowledge that this thinking is circular).
  3. Similar to BTC, ETH carries a very strong Lindy Effect as the first collateral in DeFi. History tells us that the adoption of a new technology/asset/product is extremely path-dependent. Oftentimes having better distribution is more important than having better product/tech. If Ethereum and DeFi do become the financial layer of the future, ETH may very well remain as one of the major collaterals because it was the first collateral at scale and the DeFi ecosystem was built around it.

All this being said, it is believable that ETH can take 10% of Bitcoin’s market share if Ethereum and DeFi continue to grow. If we take the potential market cap of Bitcoin to be around $4.7 — $14.6 trillion, then ETH’s potential monetary value could be at $0.5 trillion — $1.5 trillion.

Potential Valuation of ETH

Ethereum has a complex narrative and mapping out the various value accretion mechanisms of ETH is even more challenging. We believe the potential value of ETH should be the sum of its valuation as a 1) consumable commodity, 2) capital asset, and 3) monetary value (subdivided into its valuation from its payment and store-of-value functions).

Our thought exercise above puts the total potential valuation of ETH between $3.7 and $4.7 trillion in the distant future base on the following parts:

  1. ETH’s valuation as a consumable is likely bounded given its high velocity.
  2. ETH’s valuation as a capital asset with cash flows is likely in low single-digit trillions. $3.2 trillion was our point estimate extrapolated from a somewhat conceivable success case.
  3. ETH’s monetary value can potentially be $0.5 to $1.5 trillion. We assumed that the value from ETH’s payment functionalities will be nonexistent and that all of ETH’s monetary value comes from it being a non-sovereign store-of-value driven by DeFi.

Important: These numbers are by no means precise estimates. We went through this thought exercise to help readers grasp the potential impact of the Ethereum network and the corresponding value of ETH. It is meant to highlight potential narratives for Ethereum and give the readers a mental model for ETH’s value accretion. We also did not probability-adjust or apply a discount rate to account for the numerous risks (e.g. technical, competition, regulation). A shrewd investor needs to take these risks into account to ultimately gauge the risk/reward of ETH as an investment.

Acknowledgement: Thank you to Haseeb Qreshi, Ashwin Ramachandran, Tom Schmidt, and the rest of the Dragonfly team for the feedback.

Disclosure: Dragonfly Capital is an investor in BTC, ETH, Matter Labs and many of the Decentralized Finance assets discussed in the essay.



Celia Wan
Dragonfly Research

Certified paradigm shift identifier because I read Kuhn thrice; History and Philosophy of Mathematics @UChicago