A High Level Overview on L2's

15 min readFeb 12, 2023

The Layer-2 Investment Thesis: Optimistic Rollups Rule the Medium Term

Setting the Stage: 2021 and 2022

In 2021, the development of a robust application layer on top of Ethereum was well underway, with DeFi at its centre. NFTs sparked mainstream attention and adoption of Ethereum soared.

Infrastructure projects that were designed to serve as foundational pillars for future applications also began to take shape. Projects including Covalent, Akash, IPFS, Livepeer, The Graph, Arweave, have made it easier for developers to build full width applications.

Late 2021 onwards, we saw significant development and successful adoption of rollups. While the scale and speed that these solutions bring are still not yet fully realized, it is safe to say that they found strong market fit in 2022.

High Gas Fees: Ethereum Is Too Expensive

As DeFi and NFTs gained popularity in 2021, activity on the Ethereum network increased (drastically increasing the demand for blockspace) making Ethereum too expensive to use for the average individual, and by extension, too expensive to build on (experiment on) for the average resource allocated team.

Gas Fees on Ethereum were as high as $200/tx during 2022

Between 2019 and 2022, a few key points became evident:

The high cost of using Ethereum, both for users and developers, hindered its mainstream adoption. The average daily transaction fee in 2020 was 10 times higher than in 2019, and in 2021 and 2022 it was 152 and 75 times higher respectively. In order for Ethereum to achieve its goal of becoming a world computer, it needs to find a way to scale effectively
Alternative Layer-1 solutions like Solana, Terra, Avalanche, and Fantom were not the ultimate answer for Ethereum’s scalability issues. These blockchains either compromised on security or decentralization, or faced scalability problems of their own when faced with high demand
Layer 2 solutions such as Optimism, Arbitrum, and Polygon PoS, while still in their early stages, showed promise as a solution to Ethereum’s scalability issues by reducing gas fees for users, and promoting increased activity among developers and transactions. Layer 2 solutions are considered the best long-term approach for scaling Ethereum

Rollups Are Here And They Work

Layer 2 solutions are considered a crucial step in scaling Ethereum, as they allow for off-chain processing of transactions and computation, enabling faster and cheaper transactions. Rollup technology is an important aspect of these solutions, as it allows for multiple transactions to be bundled together and processed all at once, increasing the throughput of the network. This rollup-centric future will allow Ethereum to handle more transactions and reduce costs for users and developers, leading to increased adoption and usage of the platform.

More Economic Activity on L2s Can Drive Value To the L2 Token

Economic activity on Ethereum is expected to move from the base layer to Layer 2 solutions in the future. This includes things like trading, lending, borrowing, Web3 Infrastructure and consumer facing applications. As more of these activities take place on Layer 2, it will drive the creation of value on these solutions, potentially benefiting the native token through appropriate tokenomics. These Layer 2 solutions can provide a high level of scalability and security, making it more attractive for users and developers to use them.

Optimistic Rollups are Set to Drastically Increase Profitability

Optimistic rollups are a type of Layer 2 solution that can greatly increase cost savings for users and profitability in the near term. Upgrades like Bedrock for Optimism, Nitro for Arbitrum, and the introduction of EIP-4844 on Ethereum will see L2 costs further compressed. This cost savings can lead to increased profitability for businesses and individuals using the networks.

Modularity as a Strong Narrative

Rollups like Optimism and Fuel align with both the modularity and Layer 2 narratives. Modularity refers to the ability to break down a system into smaller, interchangeable parts that can be easily combined and recombined to create new functionality. This concept is important in blockchain technology as it allows for more efficient and cost-effective development. Rollups are a key aspect of achieving this modularity. They also align with the Layer 2 narrative, which emphasizes the importance of off-chain processing and scalability in the long-term sustainability of blockchain networks. Both of these concepts are expected to be prominent in the next two years as the industry continues to evolve and mature.

What do Rollups even do?

To put it simply, rollups bundle multiple transactions into a single block, called a roll-up block, that is then recorded on the Ethereum blockchain, resulting in increased throughput. By compressing transaction data and conducting it off-chain, rollups are cheaper as they require less storage space and computation. The data is still secured on the L1, without the need for full computation on the chain. Funds in a Rollup are held by smart contracts, and rollup operators stake funds in the L1 contract to ensure security.

Optimistic Rollups

Optimistic rollups rely on the assumption that all transactions conducted outside of the main network are valid.

If a transaction is found to be fraudulent, an operator must submit a fraud proof to have the transaction re-evaluated. Fraudulent operators will have to pay a portion of their stake to the network. Users can undo incorrect transactions and penalize bad actors by taking away a portion of their staked funds.

Optimistic rollups run alongside the main network, summarizing and verifying transactions. They offer the advantage of being fully compatible with the Ethereum Virtual Machine, allowing for the same functionality as on the main network. However, one drawback is that withdrawals from the Optimistic rollup to the main network can take several days.

Zero Knowledge Rollups

Zero Knowledge rollups work slightly differently to Optimistic rollups, in the sense that with a ZK rollup, all the data behind an L2 block header (all the data of all the transactions in a L2 batch) is not required to be posted onto the Ethereum main chain.

Instead, the L2 simply posts the state difference along with a validity proof that guarantees that all the transactions under the block header are valid. This greatly compresses the amount of data to be stored on an L1, drastically decreasing fees without compromising on security. ZKRs aren’t EVM compatible, and hence, compromise on composability.

ZKRs use validity proofs and do not assume that any transaction is valid, rather guarantee it. This reduces withdrawal times compared to those of fraud proof based rollups.

The zkEVM

The zkEVM is a virtual machine that executes smart contracts in a way that is compatible with zero knowledge proof computation. This essentially means that applications can be deployed onto ZK-Rollups, with more or less the same codebase they use for Ethereum, preserving composability with the EVM while not compromising on security. zkEVMs weren’t expected to be ready until late 2024, but over the last year, the pace of the entire zero knowledge proof ecosystem has exceeded even experts’ expectations.

zkEVMs are still in their initial testing phase, and will likely see significant battle testing through 2023. Scroll, zkSync, Polygon have all launched zkEVMs running in testnet.

State of Layer-2 Networks: Optimism, Arbitrum, Polygon PoS, Starknet, ZkSync

Transaction Activity

I will present three charts to provide an understanding of the development in terms of network activity in the L2 space. These charts should offer an understanding of the current state of activity across L2 solutions, the strength of these solutions relative to Ethereum, and where I anticipate seeing significant growth in activity in the near, medium and long-term.

Short Term: Polygon PoS and Ethereum

Polygon has been the clear leader in terms of network activity among L2s. The network gained widespread attention in early 2021 when it effectively alleviated congestion on Ethereum.

Currently, Polygon processes 3–4 times more transactions daily than Ethereum. This success is largely attributed to a strong business development team that has formed partnerships with well-known brands such as Reddit, Nike, and Disney, attracting significant amount of retail attention. Additionally, the Polygon network also provides a platform for Web3 games, which tend to have high transaction frequency.

It is expected that in the short-term, Polygon PoS will continue to dominate EVM-based network activity, while the team works on developing its Zk-oriented L2 solutions.

Medium Term: Optimism and Arbitrum

Arbitrum and Optimism, both optimistic rollups, were launched in mid-late 2021 and have seen rapid growth in transaction activity.

Together, they now process more daily transactions than Ethereum. This growth was likely driven, in part, by speculation and anticipation of token airdrops. However, Optimism’s network began to gain traction in the latter part of 2022, well after its token launch.

Arbitrum and Optimism have seen ATH usage on the network through 2022

In 2023, it is expected that this growth will continue to increase due to network upgrades, the expansion of their ecosystems, speculation surrounding Arbitrum’s token launch, Optimism’s “Bedrock” upgrade and improvements in the base layer scaling. By the end of 2023 and leading into 2024, it is likely that Optimism and Arbitrum will surpass Ethereum in terms of daily transaction activity.

Long Term: Zk Rollups & zkEVMs

While zk scaling solutions have been around since June 2015 (zkSync) — we have seen slight adoption only begin over the last year or so. The lack of composability is probably the main prohibitive factor to serious adoption (1m+ daily transactions), and hence, dapps have found it much easier to deploy on optimistic rollups or sidechains, and the users follow.

The technology used in Zk Rollups (generating validity proofs among other things), is rather complex, and optimization along with scaling of this technology will continue to take time. However, once fully developed, ZK EVMs could be the endgame for Ethereum scaling.

Starknet and zkSync have the lowest tx activity across L2s

2023 will be a big year for ZK Rollups with the zkEVM bringing composability into the equation. However, it is more or less going to be a year of battle testing with tight guardrails.

Users and User Growth

When it comes to users, Polygon dominates the metric.

With over 217m cumulative wallets (even discounted by 50% is ~100m users) since launch, Polygon has ~100x more the users than Optimism and Arbitrum, each of which command just a 1% market share. This once again, can be attributed to Polygon’s strong business development team, a much earlier launch, a rich NFT and gaming ecosystem in addition to a strong DeFi ecosystem, producing more consumer-friendly applications on the chain.

Polygon Controls ~98% of the L2 user market

Polygon: 217.6m total users
Arbitrum: 2.65m total users
Optimism: 2.66m total users

While Polygon may have gotten ahead because of an early mover advantage (at the right time), zooming into the last couple of years really gives us a sense as to where the momentum might be shifting.

In 2022,

Polygon saw declining user growth as alternate L1s (Solana, Avalanche) and L2s began eating away at its market share. Users fell -41% in 2022, compared to 2021
For Optimism and Arbitrum, the case was quite different. Users on Arbitrum grew ~5x in 2022, while users on Optimism grew ~14x
There is still significant room for these rollups to acquire market share, and it is likely that we see high-multiple user growth continue in the medium term

TVL and Bridged Value

Periodically, the total value locked (TVL) and bridge activity on Layer 2 solutions can experience sudden increases due to speculation surrounding token airdrops for the L2s or projects built on these L2s. However, over the long-term, TVL and bridge activity are reliable indicators of the economic demand for a particular chain.

Currently, there is approximately $6.5 billion in total locked value across Layer 2 networks, including sidechains, optimistic rollups, and zk rollups. Together, L2s make up just 22% of Ethereum’s TVL, and there is potentially quite a bit of value left to capture.

Arbitrum Dominates L2 TVL, still only 22% of Ethereum TVL

Arbitrum leads L2 TVL — with ~42% of total, with Optimism and Polygon the next biggest
Even in terms of total ETH bridged from Ethereum to L2s, Arbitrum is miles ahead with over 2 million ETH in bridged value

Both of these metrics suggest that Arbitrum currently has the highest level of economic user demand (dApps, token airdrop farming, trading, NFTs), as evidenced by the highest value flows to the network. To put it simply, there are more applications on Arbitrum that users want to interact with than any other L2. It is likely we see stiff competition from other L2s like Optimism in 2023.

Daily Smart Contracts Deployed

Smart contracts deployed on a blockchain is a good metric to gauge developer interest in a L2. The more smart contracts deployed on an L2 — the higher the experimentation, the more apps that get built, the more users can interact with the chain.

Polygon dominates, once again highlighting their strong BD team

While Polygon dominates the metric, looking deeper into the two Optimistic Rollups, smart contracts deployed have been steadily increasing on Optimism and Arbitrum.

In 2022, Optimism and Arbitrum saw an average of 28–30 contracts deployed daily. As both the rollups become cheaper after base and execution layer upgrades, and as they roll out grants and development incentives (Optimism is already incentivizing ecosystem growth with its retroactive public goods funding initiative), it’s likely that we’ll see this number increase 2–3x.

Fees

Rollups have found success in making transactions more affordable for average users by decreasing transaction fees compared to the L1. Optimism and Arbitrum can further decrease fees with upgrades like EIP-4844, and protocol-level improvements like Bedrock for Optimism. However, a challenge that rollups face is the bottlenecking of their execution layers. When there is high demand for their execution layer, gas fees on the chain will increase, which can negatively impact the cost savings for users. One potential solution for this is the use of modular stacks.

Overall, L2 scaling solutions, including rollups, provide significant cost savings for users and are generally interchangeable for one another, as paying a small difference in fees for a transaction or a swap is insignificant. What will end up being the key differentiators for these rollups in the medium term is how effectively they can bootstrap their respective ecosystems and provide an easy route for developers, while offering a strong UX for users. Simplicity will be key.

It costs $2.58 in gas fees to swap a token on Ethereum, while on an L2, it costs between $0.08 and $0.30 to swap tokens, a near 20x cost savings.

Rollup Economics

It is pretty clear from the core fundamental metrics (transaction activity, cost savings, contract growth, developer activity, user growth, TVL) that rollups are definitely the future of Ethereum scaling, and their development over the last year has been more than encouraging.

Other factors to consider while looking at L2s from an investment standpoint, are

How much value they can generate (how profitable are they)
How is that value distributed (how do token holders/ecosystem benefit)

Costs

Rollups have expenses, both fixed and variable, that are necessary to keep them running. These fees are paid by the rollup operator to the main Ethereum network, and passed on to users through transaction fees. The costs include: settlement costs, state write costs, data availability costs and proof costs.

Currently, Rollups pay around $1.29m/month (on average) in costs to Ethereum, which is ~70% of the revenue they generate. These costs have been decreasing over time due to protocol upgrades and a decrease in network activity on Ethereum.

The majority of the costs (~90%) are fixed costs for data availability, which is expected to decrease significantly with the implementation of EIP-4844, making rollups significantly more profitable in the short to medium term.

Fees Earned

The way current rollup economics are structured, optimistic rollups amortize costs paid to Ethereum across all the transactions in a batch, and preemptively charge users this fee as part of the transaction fee.

Arbitrum and Optimism have earned a combined $65.8m in transaction fees since they launched. Fees have trended downwards more a result of a decline in ETH prices owing to the current bear market and tight economic environment, but recently we have seen monthly fees in an uptrend.

Revenue (Profit)

L2s charge a small markup on the transaction fees for users (an L2 surge charge), that accrues to the protocol as revenue.

Currently, both Arbitrum and Optimism have an average profit margin of ~29.1% which is quite low given the valuations these chains command. However, with increased activity and cost compression over time, rollups are expected to become far more profitable.

EIP-4844

Rollups have reached the minimum level of fees they can charge for transactions and to lower the cost, they need to decrease the cost of posting and retrieving data from Ethereum. EIP 4844, which utilizes “blob carrying transactions” instead of “calldata” is expected to significantly lower the cost of operating rollups in the short to medium term and pass on cost savings to users.

Layer-2 Token Models

So far, Arbitrum and Optimism have been using a single sequencer to order and execute transactions, produce L2 blocks and post transaction data to Ethereum. These sequencers are controlled by centralized entities (Offchain Labs, Optimism Foundation) as the rollups slowly look to shed their training wheels. The next step is progressive decentralization of the rollup.

Since these rollups charge users fees in ETH, there is no native fee utility to an L2 token so far (Polygon is not considered). Decentralizing the sequencer (and hence decentralizing block production) introduces a value accrual mechanism (and demand sink) for the native token.

If you assume that eventually, rollups will be the hub for major activity across apps, then rollup blockspace becomes extremely valuable.

An L2 token can capture this value by tokenizing and selling the right to collect fees as a block producer
In a PoS like model, rollup operators can stake the L2 token and win a right to produce L2 blocks, collect transaction fees from those blocks, and emissions-based rewards (if any)
Regular token holders may delegate their tokens to rollup operators and earn a share of the fees collected
MEV will also become a larger part of the L2 block production process, and as L2s and their ecosystems mature and get more complex, there will be higher value MEV transactions that will eventually flow towards those that have the right to order transactions and produce blocks (sequencers), increasing fee capture potential

So far, Optimism (OP), Boba (BOBA) and Metis (METIS) are the only optimistic rollups with a live token, and of them, only Optimism seems to have meaningful activity on the network. Optimism’s tokenomics are yet to be fully explored, with the distribution of the token community oriented, however, yet to have any meaningful utility.

It will be interesting to see what kind of token models projects like Arbitrum, Fuel Network, zkSync, Scroll implement when they do launch their token. One thing is clear; the token should be able to accrue value from the demand for network usage.

Conclusion

In the Ethereum Layer 2 scaling landscape, Optimistic Rollups are viewed as a potential medium-term solution, with Zero-Knowledge Rollups (ZKRs) currently awaiting testing and a training wheel launch. The L2 ecosystem is forecasted to experience continued growth, capturing a larger share of Total Value Locked (TVL) from Ethereum in 2023 and 2024.

As projects such as Arbitrum and Fuel launch tokens, the tokenomics of rollups are expected to become more defined, leading to increased experimentation with Layer 2 design and architecture. The implementation of EIP-4484 is anticipated to significantly reduce rollup costs in the near future. Modularity is emerging as a noteworthy trend, however, the approach to decentralized resources remains to be seen. I see Optimism (OP) and Arbitrum (ARB) outperforming ETH as the next bull run approaches while, in the medium term, being able to capture a substantial portion of economic value from Ethereum.

This post is just a summary of some of my research into L2s building on Ethereum, sprinkled with some hopium.

Not financial advice.