「EN」ABCDE: Rethinking of 3 hot Tracks in HK

April 11th, 2023 | Lao Bai

Recently, the conference in Hong Kong has been in full swing, and the primary market has rebounded. In Q1, ABCDE team has Looked at more than 100 projects and tried the water temperature. There are several tracks that are particularly hoy, including Appchain, ZK, and Gaming.

Since this year, there have been some changes in these 3 tracks, and we have also had a rethinking and recorded our thoughts as a memo for your reference.

I. Appchain(especially RAAS, Rollup as a Service)

RAAS is a track that has emerged in the past two months, closely related to the release of OP Stack. However, Appchain as a Service has existed for a long time, with Cosmos SDK as the representative, and it has gradually become popular after Celestia proposed modular blockchain. RAAS can be seen as one of the recent hot subsets.

Why is the Appchain as a Service track likely to be overhyped recently?

First, if you are a developer and want to create your own Appchain, your options are as follows:

If your chain is EVM-based, then you can:

1. Create a pure ETH sidechain similar to Ronin (which is unlikely nowadays)
2. Use Skale to create an ETH sidechain
3. Launch a chain using Avax and connect to the Avax P chain
4. Create an EVM chain using the Polygon Supernet
5. Build a sidechain based on the BNB Chain using BAS
6. Build a Rollup Appchain using OP Stack
7. Build a Rollup Appchain using Caldera (which is essentially OP Stack)
8. Build an L3 using Zk-Sync (expected to be seen this year)
9. While you’re working on it, Arbitrum has also released their Orbit, which is a similar L3 infrastructure to Opstack
10. There are still many projects under development such as Opside, Stackr, and Sovereign SDK

If your chain is non-EVM-based, then you can:

1. Use the Cosmos SDK to build a chain, either independently or share the security of ATOM (the ICS proposal has just been approved)
2. Launch a chain using Substrate, either auction for a Polkadot slot, connect to the Octopus Network, or do not connect to anything and purely on your own
3. Build a Rollup Appchain using Celestia’s Rollkit, use Celestia for data availability, and choose settlement
4. Build a Rollup Appchain using Dymension
5. Build a Rollup Appchain using Saga
6. Build an L3 based on Starkware (expected to be seen this year)
7. There are certainly many others under development that I am not aware of or have missed.

Don’t you feel like “the options are a bit too numerous?!”

Secondly, which applications are suitable for being developed into an Appchain? Do you remember the article and debates online a few months ago about whether Uniswap should become its own Appchain?

Overall, if Uniswap were to become an Appchain, it would have advantages such as transaction fees, token value capture, resistance to MEV, and resource monopolization. However, it would also suffer from the debuff of reduced user experience, security, and composability.

In any case, we have not seen any plans or “ambitions” for Uniswap to become an Appchain yet.

Compound originally wanted to use Substrate to create its own chain, but later gave up. Currently, both Compound and Aave V3 are deployed on multiple chains, and it seems unlikely that they will become Appchains.

As for Curve, it never had such plans to begin with.

Only DYDX, which does not heavily rely on composability, has chosen to create an application chain on Cosmos. It should be seen in Q3 of this year. After Luna, this is the most anticipated Appchain, and it will show many builders a path — if you dominate a race track and do not require high composability, then it makes sense to create an “autonomous + high-performance” Appchain. Before that, you can choose a general chain ecosystem to develop your Dapp and upgrade your skills, and then fly solo when your level is high enough.

This path is relatively feasible. For example, if there is a SocialFi DAPP based on Lens Protocol that becomes popular with high daily activity, and Polygon’s throughput cannot keep up, then it makes sense to create a Rollup Appchain using the infrastructure mentioned above. However, in the short term, there are still not many projects that can meet the conditions for developing a chain using this case and the number of infrastructure components that can launch a chain.

Another option is the race track that is naturally suitable for launching an Appchain directly. The project that takes this path would be the biggest user of the aforementioned infrastructure. Currently, before the problem of asynchronous composability is fully resolved, it seems unlikely that Defi projects outside of the Cosmos ecosystem will take this path. The most suitable race track is undoubtedly Gamefi (including Onchain Autonomous World). For example, DFK and Crabada based on AVAX, OPcraft that was popular for a while, and Curio using Caldera to launch its chain.

Looking at my own Metamask and Keplr Wallet, apart from DFK and Osmosis, I haven’t used many other Appchains since Luna’s collapse. This makes me wonder if we really need such a large infrastructure of “dozens of Appchains or Rollups as a Service” in this cycle？

Regarding the recent popular RAAS (Rollup as a Service), a slight “Roast” can be made separately.

1.1. OP Series

Currently, the RAAS based on the OP series is basically Fork OP Stack. There is no technical barrier in itself, and the code and documentation of OP are very well-organized and clear. Our ABCDE technical experts can set up an OP Rollup Appchain following the documentation in less than a day. Therefore, this type of RAAS provides more value to customers in terms of Sequencer, block explorer, and fast deployment, and its marketing ability is far greater than its technical ability.

In addition, the TPS, block generation time, and usage fees of these Appchains are basically the same as Optimism, without any performance optimization measures. Therefore, theoretically, unless OP is congested, you cannot get a “better experience” than the OP general chain on this AppChain. This type of RAAS also inherits all the current drawbacks of OP, such as the fraud proof that has been delayed for a long time and has not yet been launched.

However, I still want to praise the vision of the OP Stack super chain. The more Rollup Chains that are deployed based on OP Stack or OP Stack Fork in the future, the more likely the OP super chain will become a “Polkadot-like” architecture that does not require slot auctions, has its own communication protocol, and is completely asynchronous and composable. Although this vision is still far away, the potential is enticing. OP Stack’s narrative has also enabled it to compete evenly with Arbitrum’s thriving Defi innovation. While I was writing this research report, Arbitrum also released Orbit. I knew that Arb could not ignore such a big piece of cake, and indeed, they launched their own token and new project at the same time. Now, we can see a competition between OP and Arb again!

1.2. ZK Series

In theory, the RAAS (Relay-as-a-Service) of the ZK series has the potential to improve the user experience of Appchains. This is because projects like ZK-Sync and Scroll, which follow the ZKEVM route, prioritize compatibility and may sacrifice some efficiency in their circuit design, making it difficult to optimize for a specific DApp. If RAAS can provide separate circuit design or optimization for different DApps, the performance and experience of ZK Appchains will undoubtedly be better than the generic ZKEVM.

However, there are few people in the world who understand both ZK and blockchain, and most of them are concentrated in Starknet, Zk-Sync, Scroll, and Polygon. As for the ZK series RAAS seen in the market, it is basically using the open-source version of Zk-Sync Alpha to create a forked chain of Zk-Sync… When Polygon and Scroll are fully launched and open-sourced, at most they will provide customer options, like when creating a Linux virtual machine in AWS, you choose between Redhat, CentOS, or Debian.

Therefore, with no significant technological barriers, business development is still king, and it is not as mature as the OP (Optimistic Rollup) series. After all, several ZK series Rollup official versions have not been launched or fully open-sourced yet, and what can be used now are only some open-source test versions, with bugs and less smooth experiences compared to OP. It is hoped that in the future, ZK series RAAS can appear, which provides separate circuit design or optimization for each Appchain.

II. ZK

If Appchain as Service can be considered representative of modular blockchains, then this year’s two major trends in blockchain — modularity and ZK.

ZK’s abstacle can be seen in the following three aspects:

2.1.Scalability

Needless to say, several major ZK-Rollup mainnets went online this year, but after going online, there were still quite a few problems.

• Completeness — currently, whether it’s ZK-Sync, Scroll, Polygon, the mainnets that went online are all EVM-compatible, except for Starknet, but there’s project called Kakarot which is very close to Starknet foundation doing ZKEVM. From ZK experts in the industry, I learned that the mainnets of several major ZK-Rollups that went online were more or less rushed to completion, and the completeness or maturity of each product has not really reached the level of “mainnet completeness” in the traditional sense. After going online, there will inevitably be various performance or bug issues, which will most likely require continuous upgrading and patching in the future. This can be seen from the delayed testing plans of various companies from last year to this year. The reason behind this is that ZKEVM is indeed “too difficult”, so difficult that even the top engineers in the industry need much more time than originally imagined to overcome it. As for why they all rushed to launch the mainnet this year, I think it’s related to the pressure brought by the continuous enrichment of the OP ecosystem, the continuous upgrading and stabilization of the mainnet, and the fact that if the ZK series don’t go online, they’ll be left behind. As long as it’s basically usable, we’ll go online first. Later, we can update and iterate as needed.
• Performance — The performance of ZK Rollups is less than or equal to Optimistic Rollups in the current stage. However, from a user perspective, the difference may not be significant because both can provide confirmation within a few seconds on the Sequencer. ZK proofs can take 10–20 minutes to complete a block, but users may not care about the “Finality” of transactions on L1 and may not notice it. The popular circuit optimization or hardware acceleration can speed up the proof time, but it does not affect the user experience.
• Cost — Optimistic Rollups can prove fraud without spending much money, while ZK proofs require a lot of computing power, which costs money. You could argue that ZK Rollups use less data than Optimistic Rollups, which means less gas cost to transmit to CallData, but the reduction in gas cost is unlikely to outweigh the additional cost of the Prover. Moreover, after the release of Blob 4844, the cost of uploading to L1 will significantly reduce, and the cost advantage of Optimistic Rollups will become even more apparent.
• Security — This is a subjective issue. Traditionally, ZK Rollups are considered more secure because they rely on mathematical proofs, while Optimistic Rollups rely on economic games. Math > Game, so ZK Rollups should be safer than Optimistic Rollups in the long run.

However, in the current stage, this may not be the case.

Security means that transactions must truly “end” on ETH L1. Currently, Arb submits transactions every two to three minutes, while Optimistic Rollups take about 10 minutes. ZK Rollups typically take 10–20 minutes to prove, and this is in the case of a block being full. If the ecosystem is not prosperous enough, and the block is not full, the time will be longer.

Therefore, although math is indeed greater than the game, the time it takes to complete transactions on ZK Rollups is still higher than that of Optimistic Rollups in the current stage. ZK algorithms, circuit optimization, and hardware acceleration need to continually reduce the time it takes to prove transactions. If one day, the proof time can be reduced to the level of 10–20 seconds (in 5–10 years?), then ZK Rollups will definitely surpass Optimistic Rollups in security.

2.2.Middleware and Others

This is actually a field that many people are relatively more optimistic about than scalability because using ZK for scalability is a very “heavy” activity, as can be seen from the fact that several Rollups have been tested for years before the mainnet is launched. Middleware is relatively lighter and can perfectly leverage the features of ZK.

The hottest middleware field is naturally Interoperability, which uses ZK proofs to eliminate the need for third-party witnesses, greatly improving the security of bridges and even connecting ecosystems that were originally disconnected or difficult to connect, such as between various Layer2s or EVM and Cosmos’s IBC. Currently, there are several teams working on it, such as Succinct Labs (which just released their Telepathy product a few days ago, implementing ETH’s one-way ZK bridge), Electron Labs (who first proposed the concept of ZK-IBC), and Polyhedra (ZK Bridge, ZK DID, ABCDE’s first portfolio).

Although this field is much lighter than Rollup, it is still a super hardcore and time-consuming field. It is unlikely to see a bi-directional, satisfactory and secure ZK bridge this year. For the interoperability completely based on ZK, we still need to plan for a timeline of 2–3 years.

As for other ZK fields, there seems to be a bit of an explosion this year, as those who have been to ETH Denver know. It can be said that there are teams using ZK for everything, such as on-chain safes, DID, oracles, and even AI and machine learning. Some of them do make sense, but there are many that make you wonder — it seems that this can also be done without ZK… In short, there is a bit of a 2017 ICO hype feeling, using blockchain as a hammer and looking for nails everywhere. At that time, there were all kinds of decentralized Uber, decentralized Airbnb-like projects that now seem very absurd, but this time, ZK has replaced blockchain, and it feels like we are trying to transform every field with ZK.

III. Gamefi

Gamefi may be the hottest track at present, without a doubt. If we count all the projects we have seen and discussed in the past few months, Gamefi should have the most sub-tracks.

The reason is not difficult to understand. On the one hand, Vitalik said years ago that finance and gaming would be the two scenarios where blockchain would first land.

On the other hand, the progress made in Mass Adoption in various directions such as Defi, storage, and Socialfi is still not ideal. Although the myth of X2Earn has been shattered, Axie and StepN have given everyone in the circle hope for “breaking the circle”. Many people firmly believe that “games” are still the key to achieving large-scale breakthroughs.

Therefore, for a while, game developers from Web2 began to turn their attention to this area, including technical teams from big companies and well-known game studios. At the same time, people from various native ecosystems such as NFT and Defi in Web3 began to consider how to “wrap their own things with Gamefi”. The Dookey Dash project by Ape was a relatively successful attempt.

However, the current Gamefi track is actually in a slightly awkward period. The pain and death spiral of X2Earn have just ended. People do not know how to strike a balance between incentives and gameplay. Everyone is blindly feeling their way forward. The consensus reached at this stage is basically “Free 2 Play”. The model of buying NFTs before playing, such as StepN and Axie, is no longer popular.

We currently see the following types of exploration:

• 3A level — going from one extreme to another. If Axie emphasizes Earn, 3A emphasizes Play. The degree of emphasis varies. For example, some that mainly focus on the Web3 circle attract Web3 players to play through NFT and other forms, while others that mainly focus on the Web2 circle use the traditional Internet mobile game model from production to operation to promotion. They just move the transaction system to the chain and may even have an embedded and seamless wallet…
• Casual social class — We have experienced the era of stealing vegetables, farms, and parking spaces in Web2. Will there be such a period in Web3 Gamefi? Social, casual, plus a little bit of Earn? It’s uncertain about the future, but at least it is a direction that has not been explored much before.
• New X to Earn mode — The most commonly seen one currently is the Bet to Earn mode represented by PSI, also called Risk to Earn. In short, it associates Earn with your skills and proficiency. Imagine a Web3 PUBG-like game, where 100 players each pay $1 to participate, and the winner takes a $100 prize. This can basically solve the Ponzi model that relied on the rapid growth of new players and the inevitable death spiral later, because this economic model has become PVP. However, similar to 3A, this model requires enough playability to retain players, as the economic incentive for ordinary users has become very small.
• “Free to Own based on NFT — Represented by DigiDaigaku, APE’s Dookey Dash also has a similar feel, which attracts players to own NFTs for free or at very low prices and then continuously empowers NFTs. This requires a high level of suspense, mystery, and marketing ability for the team in the early stages, as well as high game development and production capabilities in the later stages. Therefore, it is a highly challenging path to take, especially since the number of NFTs is often limited to 10,000 in the early stages, and how to expand the circle of people is also a significant challenge.
• Nintendo Games — This is naturally represented by TresureDAO and Gala, with Gala’s games being a bit more “heavy” and TresureDAO’s games being more focused on “mini-games”. After the success of Beacon, it will undoubtedly attract more similar mini-games to enter its ecosystem. Mini-games have always been popular on Web2, and no one knows whether Web3 can successfully replicate this.
• Defi gamification — Here DefiKingdom is unique, with very ambitious and complex systems that give people the feeling of “Chain-based Journey to the West”. However, after the 100x hype in 2021, there has been little progress, and the coin price has remained sluggish after returning to its original point. This type of complex Gamefi deeply integrated with Defi seems to have a rough future.
• Fully Onchain Game — This may be the most hyped Gamefi category at ETH Denver, for no other reason than the other categories all carry some shadow of Web2.5, and only fully on-chain games are completely Web3 games with blockchain characteristics. Some may not even be called games but instead called Onchain Autonomous World. This may be the third “blockchain-native” product after Defi and NFT. However, just as Defi (MakerDAO) and NFT (CryptoKitties) were born in 2017 and exploded in 2020 and 2021, Fully Onchain Game is also in an extremely early exploratory stage, and it may take 3–4 years to achieve a high point.

It is challenging to say which one or several of these seven directions can make it to the end. Moreover, the development of Gamefi still faces a fundamental contradiction. Although games are the easiest way to bring users from outside the circle, the essence of games is to create a world that is disconnected from reality, allowing people to temporarily escape from their busy work and life and enter the “pure land” of the game paradise. However, the connection with Web3 or the underlying blockchain inevitably makes Gamefi reconnected with the real world through some form of financialization. Whether this connection will break the most fundamental function of the game, the “pure land,” is a problem worth pondering.

Summary:

After rethinking these 3 hottest trackw, it does not mean that we are not optimistic about these tracks. On the contrary, ABCDE is optimistic about the above three tracks in the long run and has already or plans to lay out in these tracks. We tend to overestimate the short-term value of some technologies, while underestimate the long-term value of some technologies. These tracks may not provide short-term investment opportunities or returns. However, for us, a long-term investment-oriented VC with a 5-year exit period, time will be our best friend.