IOSG: Market Cap in the tens of billions or even hundreds of billions of US dollars, is the Rollup token overvalued or undervalued?

Background

Currently, the Ethereum Rollup L2 ecosystem is taking shape, with a total daily TVL of over $37b, which is more than 3 times that of Solana and over 1/5 of Ethereum’s. From the user’s perspective, the recent average daily user count of mainstream L2 has reached 158k, exceeding Solana’s data of around 100k.

However, the short-term performance of Rollups’ token price is not as expected. In terms of market cap, among mainstream Rollups, Arbitrum has a market cap of $7.8b, Optimism has a market cap of $7.3b, Starknet has a market cap of $6.9b, zkSync, which just completed the airdrop, has an FDV of $3.5b, while Solana’s FDV reached $74b during the same period. The recent launch of zkSync and its relatively poor market performance also failed to meet the market’s expectations for Rollups.

From the perspective of revenue, Ethereum’s revenue in 2023 reached $2b, while Arbitrum and Op Mainnet, which performed well in the same year, had annual revenues of $63m and $37m, respectively, with a significant gap compared to Ethereum. Base and zkSync, which performed well in the market this year, respectively achieved revenues of $50m and $23m in the first half of 2024, while Ethereum created $1.39b in revenue during the same period, and the gap has not narrowed. Rollups currently cannot achieve the same revenue scale as Ethereum.

The current low activity of some Rollups is certainly one reason, which is a problem faced by most public chains. What we are more interested in is how well Rollups have completed their mission as Mass adoption infrastructure, and whether their value is underestimated due to the current low activity.

It all comes back to the earliest proposition, the birth of Rollups originated from the increasing congestion of Ethereum, and the fees have reached a level that users cannot accept. Therefore, Rollups came into being with the inherent purpose of \u201cdecreasing transaction cost\u201d. In addition to the well-known security at the Ethereum L1 level, the advantages of Rollups also include its disruptive cost structure, known as \u201cthe more users, the cheaper the Rollups\u201d.

If this can be well implemented, we believe that Rollups have irreplaceable value. A more reasonable cost structure can also improve the resilience of Rollups in the face of market changes. The continuous investment brought by healthy cash flow is the source of competitiveness. Protocols with advantages in profit margins will naturally have higher valuations and long-term competitiveness.

This article briefly analyzes the current economic structure of Rollups and looks forward to future possibilities.

1. Business Model of Rollups

1.1 Overview

The Rollups protocol charges users for transactions on Rollups using the Sequencer as a financial hub to cover the costs incurred on L1 and L2, as well as to generate additional profits.

On the income side, the fees charged to users by Rollups include:

• Basic fees (including congestion fees)
• Priority Fee
• L1 related cost expenses

The potential costs that can be captured by formulating strategies on its own protocol include:

• MEV fees

On the cost side, it includes the relatively small proportion of L2 execution costs and the main part of L1 costs, including:

• DA Cost
• Verification Cost
• Communication Cost

The difference between Rollups and other L2 commercial models lies in their cost structure. The cost of DA, which accounts for the largest proportion, is considered as a variable cost that changes with the data volume. The verification cost and communication cost are more regarded as fixed costs for maintaining the operation of Rollups.

From a business model perspective, we hope to clarify the marginal cost of Rollups, that is, to what extent the additional cost of an additional transaction can be smaller than the average cost of each transaction, to verify the specific extent to which the principle of “the more users, the cheaper the Rollup” holds true.

The reason behind this is that Rollups batch process data, compress data, verify aggregation, resulting in higher efficiency and lower marginal cost compared to other public chains. In theory, the fixed cost of Rollups can be well spread out to each transaction, so it can even be negligible when the transaction volume is large enough, but this also needs to be verified by us.

1.2 Rollups Revenue

1.2.1 Trading Fee Income

The main source of income for Rollups comes from transaction fees, namely gas. The purpose of the fees is to cover the costs of Rollups and to obtain a portion of the profits to hedge against the risk of long-term L1 gas fluctuations, as well as to obtain some profits. Some L2s will charge transaction priority fees to allow users to prioritize urgent transactions.

Aribtrum and zkSync adopt the FCFS mechanism, that is, transactions are processed on a first-come, first-served basis, and do not support “queue cutting” requests. The OP stack takes a flexible approach to such issues, allowing for “queue jumping” for transactions with priority fees.

Source: IOSG Ventures

For users, the cost of Rollups L2 will be determined by the lower limit base fee when the on-chain is less active. When the on-chain is busier, congestion fees will be charged based on the congestion level judged by each Rollup (often rising exponentially).

Due to the extremely low L2 overhead of Rollups (only off-chain engineering and operation costs) and the high autonomy of the execution costs charged, almost all user income used to pay L2 fees will become the protocol’s profit. Due to the centralized operation of Sequencer, Rollups have control over the basic fees, congestion fees, and priority fees, so the L2 execution fees will be a ‘parameter’ game for the protocol, allowing for any design as long as the ecosystem is prosperous and the price does not provoke user resentment.

Source: David_c @Dune Analytic

1.2.2 MEV income

MEV transactions can be divided into malicious MEV and non-malicious MEV. Malicious MEV refers to front-running transactions similar to sandwich attacks, which aim to seize the value of users’ transactions. For example, in a sandwich attack, an attacker inserts their own transaction before a user’s transaction, causing the user to buy at a higher price or sell at a lower price, known as being ‘sandwiched’.

Instead of malicious MEV, arbitrage and liquidation, such as back-running transactions, arbitrage can balance prices between different exchanges and improve market efficiency; liquidation can remove bad leverage and reduce systemic risk, and is considered beneficial MEV behavior.

Source: IOSG Ventures

Unlike Ethereum, Rollups do not provide a public mempool, only sequencers can see transactions before they are finally confirmed, so only sequencers have the ability to initiate L2 on-chain MEV. Since most L2 are currently centralized sequencers, there will be little malicious MEV for the time being, so current MEV income will need to consider arbitrage and liquidation types.

According to the research of Christof Ferreira Torres et al., it replayed transactions on Rollups and concluded that Arbitrum, Optimism, and Zksync have on-chain non-malicious MEV behaviors. The three chains have currently generated a total of $580m worth of MEV, which is worth paying attention to as a source of income.

Source: Rolling in the Shadows: Analyzing the Extraction of MEV Across Layer-2 Rollup

1.2.3 Cost and Expenses Related to L1

This part is the fee charged to users by Rollups to cover L1-related costs, and the specific cost composition will be discussed later. Different Rollups have different charging methods. In addition to predicting and covering the cost of L1 gas for L1 data, Rollups will also generate additional fees as a reserve fund to cope with the future fluctuation risk of gas, essentially as a source of income for Rollups. For example, Arbitrum will add a ‘Dynamic’ fee, and OP stack will multiply the fee by the ‘Dynamic Overhead’ coefficient. Before the EIP4844 upgrade, this part of the fee is estimated to be about 1/10 of the DA fee.

1.2.4 Distribution

Base, due to the adoption of OP stack, has a relatively special profit sharing mechanism. Base promises to contribute 2.5% of the total revenue/L2 trading profits after deducting the cost of submitting data to L1, or 15% of the higher of the two, to the OP stack. In return, Base will participate in on-chain governance of OP Stack and Superchain, and receive up to 2.75% of the supply of OP tokens. Based on recent data, Base’s contribution to Superchain’s income is 5 ETH/day.

We can see that Base provides a significant proportion of revenue for Optimism, and in addition to cash flow, the healthy network effect makes the OP Stack ecosystem more attractive to users and the market. Although certain metrics of Arbitrum, such as TVL or the market capitalization of stablecoins, are higher than Base + Optimism, they currently cannot surpass the transaction volume and revenue of the latter. This can also be seen from the P/S ratios of the two - considering Base revenue, $OP’s PS ratios are higher than $ARB by 16%, reflecting the additional value that the ecosystem brings to $OP.

Source: OP Lab

1.3 Rollup Costs

1.3.1 Ethereum L1 Data Cost

The specific cost structure of each chain varies, but can generally be divided into communication costs, DA costs, and the unique verification costs of ZK Rollups

Communication costs: mainly include state updates between L1 and L2, cross-chain interaction, etc.

DA cost: including compressed transaction data, state root, ZK proof, etc. published to the DA layer.

Before EIP4844, the main cost of L1 comes from DA cost (over 95% for Arbitrum and Base, over 75% for zkSync, over 80% for Starknet)

After EIP4844, the cost of DA dropped significantly, and due to the different mechanisms of different L2s, the degree of cost reduction also varies, with a decrease of about 50% to 99%.

1.3.2 Verification Cost

Mainly used for ZK Rollup to verify the reliability of Rollups transactions through ZK means.

1.3.3 Other Costs

It mainly includes off-chain engineering and operation and maintenance costs. Due to the current operation mode of Rollups, the operating cost of nodes is close to the cost of cloud servers, which is relatively small (close to the cost of enterprise AWS servers)

1.4 Comparison of L2 profits and other L1 data

So far, we have a general understanding of the overall income-expenditure structure of Rollup L2, which can be compared with Alt L1. Here, Rollups have chosen Arbitrum, Base, zkSync, and Stakrnet’s weekly average data as the source of data.

Source: Dune Analytic, Growthepie

It can be seen that the overall profit margin of Rollups is quite close to Solana, showing significant advantages compared to BSC, reflecting the excellent performance of Rollups’ business model in profitability and cost management.

2. Rollup Horizontal Comparison

2.1 Overview

Different stages of Rollups development have significant differences in their fundamentals. For example, when there is an expectation of issuing coins in a transaction, Rollups will see a significant increase in trading volume, as well as a significant increase in fee income and expense.

Source: IOSG Ventures

The vast majority of Rollups are still in the early stages, and absolute profitability is not so important for them. It is more important to ensure a balance of income and expenditure and long-term development. This is also the concept that Starknet has always announced, not to charge users extra fees and to achieve profitability through this.

However, since mid-March, Starknet has been operating at a continuous loss and its on-chain activity has indeed been poor. What are the fundamental reasons for the continuous negative revenue, and will it continue in the long term?

Let’s continue to delve into this issue. In fact, the income structure of Rollups is relatively convergent, and the cost structure of marginal cost brought by the Rollup mechanism of each chain is different, and the cost differences are also brought by different data compression methods and calculation mechanisms.

Source: IOSG Ventures

We hope to compare the costs in Rollups to help us horizontally compare the different features of Rollups.

2.2 Cost Structure of Different Types of L2 ZK Rollup

ZK Rollups are mainly different in terms of verification cost, which is often considered as their fixed cost. It is difficult to offset the verification cost by charging transaction fees, which is also the root cause of the insufficiency of Rollups revenue.

Source: David Barreto@Starknet, Quarkslab, Eli Barabieri, IOSGVentures

This article mainly discusses two relatively mature ZK Rollups with trading volume.

Starknet

Starknet uses its own SHARP verification service for transaction sorting, confirmation, and block generation. After forming batches through SHARP, transaction proofs are sent to L1 contract verification, and once approved, the proof is sent to the Core contract.

The verification and fixed cost of DA in Starknet come from blocks and batches respectively.

Source: Starknet community - Starknet Costs and Fees

The variable costs in Starknet increase with the number of transactions, mainly DA costs, which theoretically do not incur additional expenses. In fact, quite the opposite—Starknet’s transaction costs are charged per write, but its DA costs only depend on the number of updated memory units, not the number of updates per unit. Therefore, Starknet has previously charged excessive DA fees.

The collection of transaction fees and the payment of operating costs are subject to time lags, which may result in partial losses or profits.

Therefore, as long as there are transactions being generated, Starknet needs to continuously produce blocks and pay for the fixed costs of blocks and batches. At the same time, the more transactions there are, the more variable costs need to be paid. Fixed costs do not significantly increase marginal costs.

Source: Eli Barabieri - Starknet User Operation Compression

Due to the limitation of computing resources for each block (Cairo Steps), the gas fee calculation method of Starknet is based on the fixed cost and variable cost, which cover the computing resources and data size, respectively. Since it is difficult to allocate the cost of block/batch to each transaction, but because each block is closed only after reaching a certain amount of computing resources (fixed cost is triggered), a part of the fixed cost can be calculated and collected through the dimension of computing resources.

However, due to the limitation of block time, if the trading volume is insufficient (the computational power in a single block is insufficient), the required amortized cost cannot be well measured by computing resources. Therefore, fixed costs still cannot be fully covered. At the same time, the ‘limitation of computing resources’ will be affected by the Starknet network parameter upgrade. The significant losses in short-term operation after EIP4844 reflect this point. The losses were not mitigated until the computational resource parameters in the collected fees were adjusted.

Source: Growthepie

Starknet’s pricing model cannot effectively fill fixed costs in every transaction, so there may be negative revenue when the Starknet mainnet is updated and trading volume is extremely low.

zkSync (zkSync Era)

After the Boojum upgrade, the zkSync era shifted from block verification to batch verification and state difference storage, effectively reducing verification and DA costs. The process is similar to Starknet, where the Sequencer submits batches to the Utor contract (state difference and DA commitment), proof nodes submit verification (ZK proof and DA commitment), and after verification passes, the batch is executed (once every 45 batches); the difference is that Starknet has verification costs for both blocks and batches while zkSync only has verification costs for batches.

Cost Comparison between zkSync and Starknet

Starknet batch sizes are much larger than zkSync Era, with zkSync Era limiting each batch to 750 or 1,000 transactions, while Starknet has no transaction limit.

Source: IOSG Ventures

In this way, Starknet’s scale ability is stronger, because each block has computational resource limits, the ability to process more transactions and batches in a single block makes it perform better in high-frequency trading and scenarios requiring a large number of simple operations, but it will encounter the problem of high fixed costs when the transaction volume is small. zkSync’s compression efficiency and flexible block resources make it more advantageous in scenarios that require flexible response to L1 gas price fluctuations and lack of on-chain activity, but there will be limitations in terms of block generation speed.

For users, Starknet’s charging model will be more user-friendly, with less correlation to L1 and stronger economies of scale. Zksync’s fees are more efficient, but they fluctuate more with L1.

For protocols, in the low activity stage, Starknet’s high fixed costs will result in more losses, while zkSync is more suitable for this scenario. In the high activity stage, Starknet is more suitable for conducting a large number of high-frequency transactions and controlling costs, while zkSync’s current mechanism may be slightly less competitive in terms of high transaction volume.

2.3Optimistic rollup

The cost structure of Optimistic Rollup is relatively simple. Without verification costs, users only need to pay for the computation cost on L2 and the DA cost for publishing data to L1. The publishing of state roots is more of a fixed cost, as it is related to block production, while the uploading of compressed transactions is a variable cost that is easier to estimate and share.

Compared with Zk Rollup, its fixed cost is lower, more suitable for scenarios with moderate transaction volume, but because each transaction needs to include a signature, the marginal cost of DA can be higher, and the advantage brought by marginal cost in large-scale adoption phase is relatively smaller.

Source: IOSG Ventures

Based on the current adoption scale, the fixed cost of ZK Rollup may result in a higher lower limit for unsubsidized transactions compared to OP Rollups, bringing costs to users, but the obvious advantage of ZK lies in scalability:

High trading volume and proof aggregation will verify cost sharing, and the marginal cost saved by L1 in the end will exceed Optimism Rollups; running Validiums/Volitions and only requiring state differences DA, faster withdrawal speed, etc. will be more suitable for the scaled economic demand and RaaS ecosystem.

2.3 Data Comparison

Income

Rollups charge users gas fees. Base has higher income, Starknet has lower income, Arbitrum and zkSync are at the same level. The difference in transaction volume leads to horizontal and vertical gaps, so we calculate the income per transaction. It is found that before the EIP4844 upgrade, Arbitrum’s income per transaction was higher, and after the upgrade, Base’s income per transaction was higher.

Source: IOSG Ventures

Cost

From the perspective of the transaction cost, before EIP4844, due to the high cost of DA, the transaction cost was high, actually in a situation of relatively high marginal cost, and the cost advantage caused by economies of scale was not reflected. After EIP4844, with the significant drop in DA cost, the transaction cost of Base has dropped sharply, and it is currently the lowest among all Rollups. Compared with OP and ZK, OP Rollups can be seen as the greater beneficiary of the upgrade. The actual cost of StarkNet’s L1 DA can be reduced by about 4 to 10 times, slightly less than one order of magnitude of OP Rollups. This is also consistent with theoretical inference: in the EIP-4844 upgrade, the benefits of ZK Rollups are not as great as those of OP Rollups. The cost performance of ZK Rollup after the upgrade also reflects the impact of fixed costs.

Source: IOSG Ventures

Profit

From the data, it can be seen that Base has the highest gross profit due to economies of scale, far exceeding Arbitrum, which is also Optimistic. Starknet, as a ZK Rollup, has a negative transaction gross profit due to the low transaction volume, unable to cover fixed costs, while zkSync has a positive gross profit but is similarly constrained by fixed costs, lower than OP Rollup. The EIP-4844 upgrade did not directly help profit margins - the main beneficiaries will be users, with their cost significantly reduced.

Source: IOSG Ventures

3. Summary

3.1 Cost Side

It seems that most Rollups are still in the front half of their Margin curve at the moment. With the increase in transaction volume, the marginal cost will gradually decrease, and the average fixed cost will also decrease significantly. However, after the rise of transaction volume in the Ethereum L1 or L2 ecosystem in the future, the increase in average transaction cost due to network capacity will lead to a gradual rise in marginal cost (as seen from the performance from Base3-5), which is an issue that cannot be ignored in the long-term development of Rollup. When paying attention to the cost changes caused by short-term adoption, we also need to focus on the efforts Rollups make on the long-term cost curve.

Source: Wikipedia - Cost curve

In the short term, the best way to establish barriers for Rollups is to reduce marginal costs more effectively, and adjusting revenue and cost models according to market conditions is a good solution.

3.2 Revenue Side

To maintain long-term competitiveness, the protocol tries not to charge users additional fees, and even subsidizes the costs to keep user expenses low and stable, as we can see in the current situation of Starknet. Priority fees can indeed bring in more revenue, but the premise is that the chain must have sufficient activity.

After EIP4844, the revenue of some Rollups has decreased significantly (such as Arbitrum), due to the near elimination of implicit revenue from a portion of the profit margin source—DA data fees. The revenue model of Rollups will become relatively single, mainly mining from L2 fees. With the increase in transaction volume, the generated priority fees and congestion fees will be important components of revenue. At the same time, in terms of active income, extracting MEV through Sequencer will also be an important source of future Rollups revenue.

Overall, the business model of Rollups does have the advantage of economies of scale, especially ZK Rollups. The current market conditions are not conducive to Rollups leveraging their advantages, and they all need to wait until a Base moment similar to March-May of this year. The diversity of business models and the adaptability of different Rollups in different market conditions also show us the far-reaching considerations of the Ethereum L2 Rollups ecosystem.

References

_5FPtPZ88