13 min readPrevious Valuation Model: We initially projected a high future price for ether (ETH), based on the assumption that Ethereum's fee burn mechanism (EIP-1559) would tie network utility directly to token value through substantial Layer 1 (L1) fees.
Impact of Layer 2 Success: The introduction of blob space in the Dencun hardfork made Layer 2 (L2) settlement significantly cheaper, which as expected, removed a demand vector for Ethereum block space. However it also, perhaps unexpectedly, encouraged existing Ethereum use cases to migrate to L2s, lowering Ethereum transaction demand as a whole.
ETH Inflation and Overvaluation: The sharp drop in L1 fee spend suggests that ETH is currently overvalued when considering our fundamental investment case model.
Possible Solutions: Restoring ETH's value accrual could involve developing new L1 use cases that generate substantial fees, achieving massive growth in L2 transactions, or repricing L2 blob space to increase their level of fee burn.
Challenges with Protocol Changes: Continuous and unpredictable changes to Ethereum's protocol economics create uncertainty, making it difficult to build reliable valuation models and potentially deterring institutional investors.
Back in the fall of 2023 we published a comprehensive investment case for ether along with a valuation model built on the back of what we identified as being the fundamental mechanisms of the Ethereum protocol that are responsible for driving the value of ETH.
In short, our thesis is that after the implementation of EIP-1559, the introduction of the burn creates a conduit for overall system utility—as measured by the purchasing power of the total fees spent on the L1 platform—to generate upwards pressure on price by creating a downwards pressure on supply. In other words, improving the utility of the Ethereum system can now accrue value directly to token holders. We thought this was quite interesting and exciting.
Taking the total fee spend of 2021 as a conceivably achievable level of fee spend 5 years into the future, we then built a model that assumed an equilibrium state between issuance and burn at a 45% stalking rate and a $10bn annual fee spend on the Ethereum L1. With those assumptions, our model put out an equilibrium price of $8,219 in 2028, if the usefulness of Ethereum was to grow sufficiently to generate $10bn of annual transaction fees by that time.
The Accrual of Value to the Ether Token is Extremely Sensitive to Fee Spend
As should be obvious from the paragraphs above, one of the major sensitivities of this model is L1 fee spend. At the time of writing, Ethereum L1 transaction fees were running at an annualised rate of about $2.5 billion, driven mostly by trading on DEXs, layer-2 projects (L2s) and transferring tokens (ETH, stablecoins, and other general ERC-20s). Among the newer entrants into the use cases generating significant L1 fee spend were the so-called Ethereum L2s. Back then, in order to settle their batches of transactions back onto the L1, L2s had to use a relatively expensive type of Ethereum function named calldata. This had two important effects:
L2s were unable to achieve the level of cost savings deemed necessary for the level of scaling they were envisioned to provide, and;
L2s generated a significant amount of L1 fees (about 10-15% of the total and anywhere from $150m - $550m annualised)
In addition, at that time, L2s had not developed far enough in their capabilities to cannibalise a significant amount of L1 gas expenditure. All in all this upheld a status quo with an ether inflation rate that tended to hover right around zero—the exact point we believe to be the most stable long-term state—and where the system seemed rather well balanced in terms of supply and demand.
But as Often Happens in Ethereum, Things Changed
When we wrote up the valuation framework, the Dencun hardfork was looming on the horizon. As one of its key changes, it would introduce the so-called blob space, which would dramatically reduce the necessity of L2s to pay significant amounts of gas. Instead of needing to settle to the L1 using expensive calldata, L2s could now settle within the blob space, and at a fraction of the cost in gas burned.
The benefits were obvious, if L2 transactions could have their costs reduced by several orders of magnitude, this would allow a huge amount of the smaller value transactions to move from the L1 over to the L2 where transaction speed is greater and transaction capacity more plentiful. The Ethereum ecosystem could also better support users of a smaller wallet size, who were seemingly being priced out paying high transaction fees and taking their business toward alternative platforms of comparable functionality (e.g. Solana, Binance Chain, etc.).
The changes in protocol economics from the Dencun hardfork looked quite spooky in terms of valuation risk, and we were clear about that in our writing. To us, the risk was as obvious as the benefit: If L2s were too successful, not only would all the calldata gas burn go away, but L2s might also attract all sorts of other transaction demand that had previously been settled directly on the L1, effectively destroying the L1 burn. If the L1 burn went too low, inflation would rise and dramatically lower the zero inflation equilibrium price point.
In other words, we feared that rather than complementing Ethereum, the L2s could become parasitic, syphoning its use cases while not paying the fees necessary to fill the leftover gap.
The Outcome, Unfortunately, was Exactly What We Feared
Turns out that pretty much our worst case scenario came to pass. The rollout of blob space made L2s so much cheaper and so much faster than the L1, that ETH supply destruction has effectively gone to zero. This has made the ETH inflation rate gradually climb, and its price has not performed well either. While the price is still nowhere near its supply equilibrium level, performance has been negative as the fee spend has dropped.
Going into Dencun the argument for why this dramatic change to the protocol economics would not have a large enough impact as to almost entirely obliterate the L1 fee spend went something along the lines of: “Decentralisation and settlement assurance matters. There are many use cases that are important enough that users will still prefer the comparatively higher security levels of the L1”. And we get that, we really do. It sounds very reasonable.
The problem is that for the things that Ethereum is actually used for, decentralisation and security seem rather unimportant for most users. And frankly, given the amount of usage that had already moved from Ethereum to Binance Smart Chain and even Solana, that should at least have been widely suspected, even if not fully appreciated.
Given the Current L1 Fees, Ether Looks Dramatically Overvalued
If we revisit our valuation model with assumptions that extrapolate the fee environment since the Dencun hard fork came into effect, we’d come to the conclusion that the market is overvaluing ETH at current prices, and that our former estimate of it reaching a unit price of $8,219 in five years time could turn out to be a miserable overcalculation.
Since L2s have received the benefit of blob space settlement in March 2024, Ethereum’s transaction fees have sharply dropped. From Q1 to Q2 2024 user spending fell from $1.1b to $480m, a relative change of negative 56%. Furthermore, spending in Q3 2024 is on pace for just $300m — a 73% drop since the hardfork activated at the end of Q1.
When we model out the months since the change we arrive at a total annual spend of nearly $1.5 billion. If we keep the same assumptions otherwise — a flat ETH supply inflation, which using the existing staking rate of 28% equates to a necessary gas price of 25 gwei — our current equilibrium value estimate for ETH is $1,541.
It’s Not All Doom and Gloom
Now, even if what we’re doing here certainly constitutes raising the alarm bells regarding Ethereum protocol economics and the negative impact we think the current gas price is having and will continue to have on the ETH price, the situation is not all doom and gloom. Let’s not forget that the reason L1 fees have cratered is that Ethereum L2s have been such a resounding success that it has literally drawn the life blood right out of the base layer.
In other words, L2s are doing exactly what they are meant to do, they are just arguably doing it too well and at too low of a cost. The gas price of the blob space was simply set too low. Unfortunately this is a bit of a typical Ethereum problem: tinkering with complicated things is extremely difficult, and it has a tendency to generate unintended consequences. But this is the general approach they have chosen, so we will simply have to live with the ongoing effects of that.
Looking at some L2 metrics, their success is undeniable. Usage is growing steadily, and the number of transactions executed per second is approaching 140, an over 10x increase above the base layer of Ethereum (13 tx/s). Given that the strategy of the Ethereum Foundation is to build as large of a network effect around the Ethereum ecosystem as possible, this approach makes sense. But what about the network economics? Absent any real fee spend, we can’t see any fundamentals-anchored mechanism whereby increasing transaction demand on the L2 should lead to any appreciable value accrual at the token level.
Three Possible Options For Restoring The Burn
As mentioned both in our original valuation paper and above, we see three possible options for restoring L1 fee spend to a level where we regain an appreciable burn rate.
Either:
A robust set of L1-specific use cases need to emerge and be willing to pay a substantial amount of gas for transactions;
L2 transaction counts need to grow astronomically (several orders of magnitude), or;
L2 blob space must be repriced
Given the current state of the Ethereum fee spend, alternative 1 seems like a distant possibility. And while L2 usage is growing very healthily, at current growth rates and protocol economics, it would take decades for L2 usage to reach the levels necessary to even replace 2023 L1 fee spend ($2.4b), let alone reach the highs of 2021 ($10bn) assumed in our model.
Unfortunately, we think the most likely outcome is alternative number 3. It’s unfortunate not because we don’t think it would be able to achieve the goal, but because it would just be another instance of the seemingly endless ongoing timeline of constant protocol tinkering, where each change on top of adding some new shiny thing tends to “fix” some previous inefficiency or unintended consequences… and then creating a whole new set of both.
Or perhaps nothing will happen, because the Ethereum community might in fact be finding themselves in between a bit of a rock and a hard place. Given that there are several (important and influential) companies now—many of which have (important and influential) Ethereum developers on their boards—who have built entire business models around the current L2 economics, the prospect of rugging these industrial partners is probably not going to seem like a great approach either.
To Tinker or not to Tinker, That’s the Question
At some point, Ethereum needs to figure out what it actually wants to be when it grows up. Yes, it is still young, and to many it is still seen as a startup kind of project—nimble and agile—pivoting as and when needed in a fast moving space. The problem with this is that Ethereum is a protocol, not a company, and protocols need stability in order to build lasting network effects. The whole point is that other people are supposed to use them for things, but that’s really difficult if you don’t know what the protocol will even look like 2 years — let alone 10 years — from now.
Moreover, as Ethereum is making an attempt at venturing deeper into the institutional investment space, Ethereans will find that professional allocators have higher expectations for fundamental investment cases than “don’t worry, we’ll figure it out later—look at our growth”; that type of approach tends to find a more receptive audience in the venture capital space.
The danger here is that, just as Ethereum had seemingly finally found a working fundamental mechanism for tying increasing protocol utility and usage to token value accrual, the underlying economics driving this mechanism are way too fluid to form a reasonable assumption set on which to build valuation models.
A large portion of the challenges that Ethereum is facing, in our opinion, comes from the unintended consequences of ongoing and major protocol changes. Thus, we suggest the Ethereum community (or Foundation) refrain from making major modifications to the protocol, at the very least as they relate to ETH economics. If the goal is an uncontrollable and open computing platform, focus on incremental changes to that end. Allow for the cultivation of long-term use cases that can add lasting value to a platform having those characteristics. Eventually, then, should those protocol features have real demand, the fees will come. Entrepreneurs will be better positioned to plan for a stable protocol, building applications that deliver on long-term promises to users, who will ultimately decide just how valuable the ecosystem becomes.
Here, and we cannot stress this enough, a culture of rapid and unpredictable changes to the protocol economics is functionally indistinguishable from no protocol economics. This may be ok for retail investors, or even HNWIs with high risk tolerance, but it will simply not fly in front of the risk committees of professional investment firms.