Validator Economics of Ethereum 2.0 — Part 2…The Ether Vacuum

Compliments of Nate Chastain and Mara Schmiedt of ConsenSys

Tl:dr: At the moment, there is a lot of chatter around the evolving role of crypto investors and how asset holders can further strengthen decentralized networks through increased participation. Currently there are multiple Ethereum based opportunities for crypto holders to put their assets to work either actively or through delegate work entities. This article will explore the theoretic profitability of a delegate work entity focused on validating the Ethereum network based on the current spec of Eth 2.0 and other various assumptions over a 5-period time window. I will provide some thoughts on the need for delegate work entities for Eth 2.0 to reach its genesis threshold of 524,288 and if the economics of Eth 2.0 at genesis provide proper incentivisation for large scale and small scale validators to participate.

A few weeks back I wrote an article on the economics of ETH 2.0 from the perspective of a small scale validator, which by definition is running one validator client staking 32 Ether either on a local machine or through a cloud provider. I promised a part 2, which is set to explore the economics of a large scale validator. If you are new to blockchain or POS protocols I would recommend reading my first article (Validator Economics of Ethereum 2.0 — Part One or Examining the Proposed Validator Economics of Ethereum 2.0) as a primer for what is to follow.

As a recap, I provided a background into Beacon Chain, thoughts on the psychology of economic risks & required rate of return, the risks associated with being a network validator and a lite competitor analysis. The article wrapped up with a validator net yield sensitivity analysis and compared the results to the risk free rate of return. After my analysis, I walked away with the conclusion that the incentive for small scale validators may be insufficient (on a pure profitability basis) given the Ethereum Foundation’s long term mission of mass distribution, unless one wants to unprofitably acquire Ether with the hope that the network succeeds (this universe of validators will certainly exist). Overall, the network ‘interest’ rate is the rate and I am not doing this to influence it — I am just here to voice my thoughts about how we get to mass distribution.

Recently, I was at a conference in London where a staking company gave a presentation on its services and the addressable market for staking. To my surprise one of the most asked questions was, “so how do you make money”? As a result, I have shifted my focus a bit for Part 2 of the validator series and will be focusing exactly on that question by providing a few different operating scenarios for a theoretical staking-as-a-service company validating based on the current spec for ETH 2.0.

There are numerous different business models that can be utilized when staking however I found that certain drivers affect profitability and cash flow more so than others.The model is a 5-period MoM analysis that includes a number of adjustable dynamic drivers. I will focus on the sensitivity of two operating drivers that I found to be the most impactful when analyzing profitability, which are:

• Validator Clients Per Node — In essence this can be boiled down to economic risk per node. How many validator clients are you going to run on each node? This decision is based on technological limitations and/or risk tolerance. For example, the Prysm Client by Prysmatic Labs specifically allows for a single node to manage multiple (maximum still unknown) of validators if the user desires.
• Ether Liquidation Schedule — In essence this can be boiled down to how much and how often a staking entity liquidates its token inflation to pay the bills. This could be a portion every month or never, it really all depends on the cash position of the firm and its risk tolerance.

The reason I chose to focus on these two are because they can be controlled (to some degree) and represent operating strategies that are independent of network dynamics, yes these decisions may be influenced by the network dynamics, specifically ether price. However, I tried to focus on how the machine works instead of price assumptions.

As always please shoot holes in what has been put together and challenge my analysis — if we can get a healthy debate started and discover new answers everyone wins.

Delegate Work Entities

A few months back, I came across a great article by Ben Sparango called Delegate Work Entities: Bridging the Gap Between Investors and Active Users. When the article was released (late 2018) the topic of new types of business models to enhance network viability and bootstrapping were only talked about and recognized in the crypto underworlds (and still really is when zooming out for a second). The majority of retail participants in crypto were/are focused on HODL’ing (which is better than increasing downward price pressure), while those in the underworlds began to think about new ways to (profitably) enhance network effects and add value to the broader ecosystem.

“At the moment, there is a gaping chasm separating speculative investment and user activity within decentralized networks and applications. The concept of delegate work entities aligns all involved party incentives through delegation of work token assets to temporarily bridge this gap until decentralized networks improve user experience. Put more simply, delegation of work allows us to bootstrap decentralized networks in a democratic fashion until a larger portion of the community is able to undertake the tasks themselves.” Ben Sparango

In June 2018, at Token Foundry we introduced the Token Foundry Standards, which is a framework for selling consumer tokens and launching decentralized networks in a way that directs the tokens offered in the sale to actual users of these networks. This was our attempt at being a hybrid delegate work entity to help empower tokenized ecosystems and decentralized projects. Below I will highlight a few of the core pillars of the standards.

• Token Buyers Must Pass an Assessment Test
• Tokens Must Be Priced Responsibly
• Tokens Can’t Be Sold Until There is ‘Proof of a Network’
• Tokens Must Be ‘Locked Until Mainnet Launch’
• Tokens Can Only Be ‘Used’ Initially
• Tokens Can’t Be Resold Until Initial Buyers ‘Prove Use’

Active Network Participation

Over the past year we have seen numerous different subcategories and flavors of delegate work entities, the most relevant to understanding the motivations of a staking entity follows the sub theme of active network participation.

Active network participation has gained traction as a method for asset holders to increase alpha by donating resources or by staking their assets to validate a specific network. The macro thesis of this strategy is recognizing we are entering the next chapter of blockchain, which will be driven by different POS consensus methodologies, where asset owners will be incentivized to validate and participate in the network, however many of them will not have the resources/time/risk tolerance to do so.

This will include layer 1 networks as well as numerous protocols built on top of these networks. All of which will have their own respective consensus mechanisms and infrastructure (validators and/or block producers) enabling significant value generation by producing blocks and/or generating consensus through staking and/or resource sharing.

This has resulted in staking-as-a-service companies sprouting up to fill the need for asset holders who wish to utilize their digital currency to earn more digital currency. For example, Staked recently raised $4.5MM in a seed round lead by Pantera Capital, with participation from Coinbase Ventures, Digital Currency Group, Winklevoss Capital, Global Brain, Fabric Ventures, Applied Crypto Ventures, and Blocktree Capital.

If you are interested in reading further content on the evolving landscape of network participation please reference a few of the articles below.

• The Evolving Role of Crypto Investors
• A Primer on Mining 2.0
• Notation + Blockchain Mining 2.0
• Defining Generalized Mining
• Messari — Generalized Mining

Validator Assumptions

Network Assumptions & Averages:

To start let’s walk through the network assumptions over the projection period as the network will determine what is up for grabs. The chart below represents yearly averages and a 5 period average to keep our analysis a bit macro in nature. The primary drivers are:

• Total at Stake (In Network) — The projection period starts at genesis, which is 524,288 Eth and ends the projection period at 3.02MM Eth, with a 5 period average of 1.43MM. A CMGR of 2.50% was utilized to yield these results.
• Eth Price — The projection period starts at $148 and gradually climbs to $854, with a 5 period average of $404. A CMGR of 2.50% was utilized to yield these results.
• Average Network Fees/Day (Eth) — The projection period starts at 600 Eth and gradually climbs to 1,727 Eth, with a 5 period average of 1,067. A CMGR of 1.50% was utilized to yield these results.

At the end of the day these variables are a black hole so I opted to sit on the conservative side of the table. In addition, you will find a network dynamics chart below that will provide a more digestible visual of the network over the projection period . Overall, I feel my assumptions for this piece of the model are quite realistic (please let me know your thoughts).

Node Assumptions & Averages:

Next, let’s take a look at the node dynamics over the projection period. I feel the most important driver here is the percentage of the total network staked. I wanted to build this out in a manner where the imagined entity does not exceed 4% market share over the projection period. To clarify this means that the other 96% of the total at stake is made up of a mix of small and large scale validators. Imagine the entity in this model is called Beacon, which has an internal diversification mandate that it cannot stake any more than 4% of the networks it chooses to validate.

• % of Network Staked — The model remains flat at 4% throughout the projection period.
• Customer Eth Staked — The projection period starts at ~21K Eth and ends the projection period at 121K Eth, with a 5 period average of 57K Eth.

It is my hope that staking enables increased decentralization vs what we see in mining today (please see the graphic below). For a deep dive into the dynamics of Ethereum mining check out, Are Miners Centralized?, which was recently published by the Alethio Team. However, as we work our way through the results of the analysis you will see that economies of scale prevails. No matter the ethos of the chain or the social movement, human behavior will result in staking at scale to earn rents.

However, as the staking ecosystem evolves and becomes more accessible than mining, staking firms will diversify risk by validating on multiple chains & protocols. In addition, it is also my belief that due to other Defi yield opportunities those who believe in Eth enough to validate will also utilize other yield opportunities such as Uniswap, Compound Finance, and Dharma. As a result, I believe that the concentration of stakers will be less centralized than mining (especially in the long term) with chain diversification becoming an asset allocation strategy in staking as different chains mature and find their specific use cases. However, I will leave you with a quote from Jonny Rhea who is developing the Artemis client on the Eth 2.0 team at PegaSys.

“Your argument that POS will not be as centralized bc staking firms will diversify is weak. I like to think of this in terms of physics. If centralization is a force, then centralized is displacement. Any amount of centralizing force will result in the same outcome. The only difference is the amount of time it takes. I want to believe that we can do better”.

Operating Assumptions:

There are a few business models that can be run when competing in this space. The largest factor that will determine the business model is whether or not you are the asset holder or you are the asset borrower (hybrid approaches are also possible). Once you get past that first determination there are a series of smaller drivers both qualitative and quantitative that determine profitability.

However, exploring the nuances of each approach is outside the scope of this article, but hopefully the community builds on this article/model in the coming weeks/months. Below you can find the primary qualitative & quantitative operating assumptions that drive the output of this model.

Node Cost Assumptions:

The model offers a cloud and a hardware toggle to help determine cost structure. For this analysis I have selected what you could call a “premium” server option, again to stay conservative on the profitability side. For the base model it is assumed that the firm is running 100 validator clients per node (3,200 Eth). It would be great to get feedback on the cost structure of this analysis from individuals in the staking or mining industry that have proper experience validating blockchain networks.

FYE Profitability Analysis — 100 Validator Clients/Node & Monthly Ether Liquidation of 40%

In this section we will look at the financial results of the model based on the assumptions laid out above. It is important to focus on the Firm based financial metrics as it represents the value accrual of the service provider net of the value passed to token holders. As seen in the chart below by charging a 12.5% fee and staking 4.00% of the Ethereum Network on the behalf of others, the operating model yields strong results over the projection period with revenues generating a 5 period CAGR of ~60%, while total assets grow by a CAGR of 67%.

Revenue: Revenue generation over the projection period starts at ~$1.6MM and steadily grows to ~$17.8MM, with a 5 period CAGR of 60.60%.

EBITDA: EBITDA generation over the projection period starts at -$325K and steadily grows to $5.8MM, with a 5 period CAGR of 77.75%. EBITDA remains negative until April Y1 (16 months).

Free Cash Flow: FCF generation over the projection period starts at -$325K and steadily grows to $4.6MM, with a 5 period CAGR of 69%. Similar to EBITDA, FCF remains negative until April Y1 (16 months).

Cash: The firm is capitalized with $1MM of cash at the beginning of the projection period and bottoms out at a low of ~$680MM in Y2 before building to $9.2MM by Y5, resulting from the ability of the firm to scale as more Ether is staked on its infrastructure. The consistent cash balances of this analysis signals that the firm could sell less than 40% of its monthly take and still remain solvent (I will leave it up to you to find the breakeven.

Total Assets: Total assets over the projection period start at $791K and steadily grow to $10.3MM, with a 5 period CAGR of 67%.

Validator Operating Results:

Sensitivity Analysis

In this section we will sensitize our two key drivers and take a look at possible outcomes. We will be stressing EBITDA, FCF, and Total Assets given changes in our two key drivers. Below you can find a recap on our primary drivers.

• Validator Clients Per Node — In essence this can be boiled down to economic risk per node. How many validator clients are you going to run on each node? This decision is based on technological limitations and/or risk tolerance.
• Ether Liquidation Schedule — In essence this can be boiled down to how much and how often a staking entity liquidates its token inflation to pay the bills. This could be a portion every month or never, it really all depends on the cash position of the firm and its risk tolerance.

Y5 EBITDA Sensitivity Analysis

Y5 Free Cash Flow Sensitivity Analysis

Y5 Total Assets Sensitivity Analysis

Conclusion

Before we close this out let’s take a look at the consumer yield at the Genesis spec. As you can see below, the yield at the Genesis spec to a small scale validator is far better than the yield generated in my Part 1 Article. However, this does not mean that your average Eth holder will stake on their own. To reach Genesis, there will need to be 16,384 validators (genesis/32), which is more than double the total Ethereum node count of 7,580. IMO even if an incredible UX were developed, this is a tall order to fill relying on small scale validators.

So what did we learn?

1. A large scale validator is able to create economies of scale that lead to greater profitability than a small scale validator, especially as the total at stake in the network increases over time.
2. Staking 4.00% of the network results in strong profitability with conservative network dynamic assumptions
3. An initial investment of $1MM results in a 5-period CAGR of 67.14%, yielding an ending cash position of $9.2MM and Eth position of $1.1MM (total assets of $10.33MM).
4. EBITDA & FCF are only negative for the first 16 months of operations and end the projection period at $5.8MM & $4.6MM respectively.
5. The return profile for small scale validators in the early days of Eth 2.0 in theory is sufficient enough to attract participants.

The model presented in this article is proof that the current spec offers solid incentives to motivate those in the crypto community with the knowledge and resources to build much needed delegate work entities. Going through this process showed me that despite the ‘lowish’ interest rate of Eth 2.0, validating the network can still be quite profitable. However, it does run the risk of centralization over time given the unprofitable return profile explored in my last article for small scale validators at the 10 million total at stake level.

I personally believe to reach the Genesis Eth threshold of 524,288 we will need large scale validators serving as delegate work entities to make this a reality. However, the USD value of Eth that needs to be staked at current market prices is $72.35MM, which is 25% of the total amount currently locked in CDP’s. This gives me a bit of encouragement knowing that by the time Eth 2.0 becomes a reality blockchain UI/UX and participation will be more mature.

I hope that you walk away from this piece with a better understanding of Eth 2.0, the importance of delegate work entities in making Eth 2.0 a reality and how the machine works for a network validator.

The economics of Ethereum 2.0 is a topic we are very interested in at ConsenSys and will continue to do our part adding value to its reality through different outlets. Please use this as an opportunity to challenge what has been presented if you disagree with it and we can get a healthy debate started.

Special thanks to Tanner Hoban, Jon Stevens, Raul Jordan, Jonny Rhea, the Alpine team, and the Alethio team for providing suggestions/feedback for this piece. Huge shoutout to Ross Canavan & Andrew Meller for providing some unreleased tracks to get me through the number crunching.

What A Time To Be Alive!

Disclaimer

Nothing in this piece should be considered investment advice.