HashKey Capital: Full analysis of the 2024 Web3 investment track

As one of the most active crypto VCs, HashKey Capital regularly analyzes and sorts out various Web3 tracks internally. On the occasion of the New Year of 2024, we will “open source” our internal track judgment and understanding as our contribution to the industry.

The authors of this article include (in alphabetical order): Arnav Pagidyala, Harper Li, Jack Ratkovich, Jeffrey Hu, Junbo Yang, Stanley Wu, Sunny He, Xiao Xiao, Yerui Zhang, Zeqing Guo.

ZK

ZK track will further expand from previous expansion, cross-chain and other scenarios in 2023 It has been expanded to more application scenarios and further differentiated into different tracks.

zkEVM

zkEVM is in type0, type1, type2 There has been some progress in several aspects. In terms of categories, type 0 is completely equivalent to Ethereum, but it still faces technical challenges such as block production speed, release and verification status due to too much emphasis on equivalence; type 1 has made improvements and compromises based on EVM, and currently overall The application experience and operation code compatibility are the most outstanding; type2 and others have launched the main network earlier and are looking for corresponding types of application development ecosystems.

Currently, specific project situations vary and need to be analyzed based on the project's own development path, such as Polygon's CDK, StarkNet's full-chain games, etc.

zkVM

The current main technical routes in the zkVM field For zkWASM, the architecture is more scalable, so in the application direction, it cooperates with exchanges to build high-performance DEX. Major projects in the zkWASM field include Delphinus Labs, ICME, wasm0, etc.

RISC V architecture direction is mainly explored by RISC0, which is more friendly to front-end languages ​​and back-end hardware than WASM, but potential problems lie in efficiency and proof time. Currently, application scenarios are also expanding, such as Reth, which simulates the Ethereum execution environment, FHE's operating environment, Bitcoin Rollup, etc.

The other is zkLLVM. =nil; Based on this technology, Type-1 zkEVM has recently been launched. Using zkLLVM, high-level languages ​​can be quickly compiled into zkSNARK circuits.

ZK Mining

In the field of ZK Mining, GPU and The efficiency of FPGA is similar at present, but GPU is more expensive, FPGA is more like prototype verification, and ASIC scenarios may gradually differentiate, such as special-purpose ASIC chips, new FHE requirements, etc.

In addition, the number of Prover DAOs has increased significantly, and computing power is its core competitiveness. Therefore, the Prover DAOs developed by the mining team will have obvious competitive advantages.

ZK Middleware

ZK middleware can include zkBridge, zkPoS , ZK Coprocessor, zkML, zk trusted computing and other verifiable computing scenarios. Among them, the ZK Coprocessor scenario is relatively clear, and each project has basically reached the test network stage; the zkML track is still relatively popular, and the progress and competition of each project have diverged to a certain extent. In addition, a new track for ZK proof sharing has also emerged (jointly sending proofs to a certain network and sharing the revenue equally after batch processing).

MEV

You can mainly focus on the early stage of the transaction supply chain, that is, the intention stage p>

There will be more and more next-generation DEX designs and infrastructure that solve LVR and improve LP, attracting more capital

Private auction/trading pools, if operated effectively, will greatly improve the trading supply chain. And the development of FHE, MPC and ZKPs deserves attention.

Currently, most systems use centralized relays, licensed solvers, trusted builders, etc. But we believe the track will eventually be permissionless to achieve the most competitive market.

MEV supply chain will have changes in APS, document execution, PEPC, etc.

OFA

Order Flow Auction (OFA) in 2023 began to rise in the year. High-value transactions will no longer flow to the public transaction pool, but to OFA, giving users back the value they create. From RFQ auctions to blockspace aggregators, there are various OFA implementations that meet various needs in terms of price discovery and execution quality. Going forward, it is estimated that an increasing number of ETH transactions will be conducted through OFA.

Blockbuilder

As we can see with relayscan , the builder market is concentrated on a few builders, some of which are high-frequency trading companies that serve their own trading needs. In the future, as the volume of CEX/DEX arbitrage decreases, the advantages of high-frequency trading may also weaken.

Relay

The relay market faces two basic problems. (1) The market is concentrated in a few companies, namely BloXroute and Flashbots; (2) There is no incentive mechanism for relays

We expect to see optimism in the future The rapid development and implementation of optimistic relays, and proposals for relay incentives.

AA

AA (Account Abstraction) track can be mainly divided into smart contract wallets and There are two major categories of modular services.

In terms of smart contract wallets, the related companies in the AA wallet track are actually similar to the pattern of the entire wallet track. It has become more and more popular to rely solely on functions to obtain traffic. It’s getting harder and harder, but it’s the wallet factory that deserves attention.

In terms of modular services, Bundler and Paymaster are functions that these basic providers must provide. In fact, these two services have become standard services.

Some current track trends include:

Most of the infrastructure has been set up, and the current development relatively stable. Judging from the data of the entire track, it has entered a rapid growth track. The number of user wallets has begun to increase since June. As of November, there have been more than 6 million Userops and about 200K MAU.

The AA development of L2 is better than that of L1, and EF is considering giving native support to L2.

The problem of DApp not supporting AA is still very serious, and there are also implementation issues of cross-chain and cross-Rollup accounts, etc. Someone needs to come up with new solutions.

Private mempool will be merged with MEV and intent to optimize user experience.

Intents

Intents has grown rapidly after receiving attention this year, despite facing some For example, there are problems such as solver evil, order flow trust, etc., but there are relatively feasible solutions.

If Intents wants to develop better, it must consider issues such as order flow and user acquisition. Therefore, from an architectural and business perspective, Intents is very suitable for MEV. The two architectures of , AA and AA are combined. For example, Builder and Searcher are the most suitable roles for matching and solver.

Telegram Bot is likely to evolve in the direction of Intent. Their advantage in order flow gives them considerable bargaining power over builders and SUAVE. This kind of bargaining The power will even exceed that of wallets larger than theirs.

DA

There are not many participants in the DA track overall, except for Ethereum , mainly including Celestia, Eigenlayer and Avail, etc., and the project progress is variable. The head effect is obvious, leaving few opportunities for the waist and tail. Projects on the DA track can mainly examine security (including data integrity, network consensus, etc.), customizability and interoperability, cost, etc. As Celestia goes online and prices rise, the overall valuation level of the DA track is pulled up. But DA is essentially a to B business, and the income of DA project parties is closely related to the quantity and quality of ecological projects.

From a customer perspective, issuing DA on Ethereum is the safest and most expensive solution. The cost of Ethereum dropped significantly after protodanksharding, so large Rollup projects will still choose Ethereum as the DA layer. At present, the customers of DA projects other than EigenDA are mainly Cosmos ecological projects, RaaS projects, etc.; while EigenDA has a special positioning, related to Ethereum but not directly related, and may be better able to attract some customers in the middle. In addition, there are some relatively early DA projects and DA for specific scenarios such as Bitcoin DA, which may gain a good market share in niche areas.

Rollup Frameworks & RaaS

The rollup market is basically saturated and awaits new development. There are currently at least 30+ VC-backed RaaS projects and infrastructure providers entering the market. There is a need to understand which use cases are successful with RaaS and which interoperability solutions can make a difference.

Some L2/L3 frameworks (such as OP Stacks) have received significant public goods funding and developer adoption.

Some specific applications, such as DePIN, may use Ethereum rollup by using a customized execution environment.

In addition, there are many new technologies about rollup recently, such as: Risc0 Zeth/Other projects can change the way rollup verifies the state without relying on validators or synchronization. Committee; when used with primitives like ZKP and MPC, FHE rollup can provide fully universal privacy DeFi, and more.

Cosmos

Cosmos Hub will continue to try to strengthen its position in the ecosystem from all aspects in the future For example, Partial Validator Security (PSS) can more flexibly allow some validators to provide cross-chain shared security (ICS) without forcing all Cosmos Hub validators to join, which can reduce the pressure on many validators and more Easy to promote; on the other hand, Cosmos Hub plans to enable multi-hop IBC to improve user experience; in terms of protocol implementation, it plans to add the functions of Megablocks and Atomic IBC to provide atomic cross-chain transactions, which can form a unified MEV market, similar to Shared sequencer, SUAVE, etc. in the Ethereum ecosystem.

In terms of the Cosmos ecosystem, the application chain development route has recently been affected to a certain extent by development frameworks such as L2, and the number of new projects has declined. However, due to its highly customizable underlying framework and strong flexibility, it can find public chain targets that have been customized and modified as the mainstream narrative continues to evolve.

Security

Safety track projects are progressing at all levels, and products include It provides tools and protocols at all levels, such as on-chain detection and interception, on-chain tracking tools, manual audit and bounty services, development environment tools, various technical methodology applications (such as fuzz testing), etc.

Each tool is more suitable for detecting specific vulnerability categories and has specific methods for checking vulnerabilities in smart contracts (static analysis, symbolic execution, fuzz testing, etc. ), but it’s still difficult for a combination of tools to replace a complete audit.

In addition to the above different positionings, the dimensions of the project inspection can also include: maintenance and update speed, vulnerability library size, carrier, actual needs of partners, etc.

AI

The current integration direction of Crypto and AI mainly includes the underlying computing infrastructure , training based on specific data sources, chat tools, data labeling platforms, etc.

Various projects in the fields of underlying computing infrastructure and computing networks have innovated at different points, but they are generally relatively early and need to be considered in addition to creating different future sustainable business expansion paths beyond these types of agents.

The data labeling platform changes the traditional manual labeling business to web3, and the ability to obtain orders is critical. At the same time, because low-threshold data annotation can easily be replaced by AI in the future, it is necessary to better focus on high-value and high-threshold data to obtain more orders.

In addition, there are many new projects that combine AI to create To C chat tools, etc.

As a Crytpo fund, we will pay more attention to ZKML, projects with data advantages in the crypto vertical field, or to C products that are closely integrated with AI, rather than different Good at large language models and other infrastructure.

DeFi

One ​​change in 2023 may be that many investors demand real yields Products such as yields preferably from LSDfi or RWA rather than emissions based yields. At the same time, as centralized exchanges face huge regulatory pressure and assets need to find liquidity, DEX has great opportunities, especially the possibility of high-performance applications brought about by the launch of L2. You can focus on the DEX opportunities on L2.

On the other hand, there are huge opportunities for projects that make Web3 yields accessible to non-crypto-native users, including institutions. Projects that abstract away on-chain elements and provide a secure environment for non-crypto-native users will bring in a lot of money.

Specifically, according to Messari’s report, the fees incurred by sustainable DEX, including well-known projects such as dYdX, GMX, Drift and Jupiter, are divided into various segments. The highest proportion in the track.

Liquidity staking continues to grow in 2023. Nearly 22% of all ETH is staked, with Lido accounting for around 32% of the ETH staking market share (as of writing). And liquid staking tokens are actually still the largest DeFi track, with a TVL of $20 billion.

Gaming & Entertainment

Studio

Studio projects have different characteristics and required attention according to their own categories, as shown in the following table:

< td>Shrapnel, Shoot City, Abyss World, SpaceNation, Nytroverse, Mythical games, Metacene

Game categories	Game items	Features	Points of concern
Casual Games	Gameta, Gomble, CoinArcade, Octo	Games and narratives with traffic and advertising are generally low-cost purchases, attracting users through the mechanism of mini-games Establish wallets and make deposits, etc., and then sell the cultivated users. Many projects will also add arena and party game modes to allow users to compete with each other and achieve a certain level of user retention.	The small game platform itself should have a good mechanism to accumulate users, a membership system or in the form of NFT
Large-scale games	Mainly client games, developed by seasoned studio, game categories are first-person shooter, MMORPG or ARPG; game development It takes a long time and the R&D cost is upwards of 10M	This type of game will pay more attention to the core producers, including track records and whether the previously produced games are in the same category; and the team needs to have strong management capabilities for the entire R&D budget< /td>
Casual/mid core	Ambrus, InfiniGods, Moonveil, Mugen Interactive, Puffverse	1, card category, tower defense category and strategy (SLG) games; usually web and mobile can support this category of games, which are easy to get started, but have a certain complexity. For players who want to go deeper, the content has enough strategy and depth to experience ; 2. The cost budget of a single product in the benchmark is around 7M	It also has requirements for producers and requires relatively strong cost control.
IP Building	SpaceNation, Aether Games, Pixelmon	Games are the main delivery products, and there are also other medium forms Producing contents such as movies, TV dramas, comics and animations can better empower the main IP. The business logic here is that some other media forms such as animation may have lower customer acquisition costs than traditional games, but the effect is more Good	The game itself can have many categories, and we will analyze it according to different categories
AI feature games	AI Arena, Oasis, Sleepless AI	Things with AI as the core are currently relatively innovative, such as AI arena. The current product is relatively simple, but the narrative can be told in a very short period of time and cannot be implemented in the short term.	You need to see whether the AI ​​feature is just a layer added to the outlet, or whether the AI ​​feature essentially improves and helps the product and users

< p style="text-align: left;">The overall game track can be simply summarized into the above categories. The overall game quality and team professionalism have been greatly improved compared to the previous cycle. We continue to look forward to the studio's entry in the future. web3 also has mature experience in product development and operation, and continues to seek game producer + KOL compound founders who have strong learning abilities, are sensitive to crypto and the community, and are willing to speak out. In terms of Web3, because the current history of chain games is short and there is not much mature experience to draw from, we will pay more attention to whether the team's ideas and ideas are in line with web3, and to learn quickly, rather than having rich web3 experience.

We will continue to pay attention to UGC in the future. Currently, web 2 UGC has problems that cannot be solved by centralization. In addition to providing tools for users to create content, UGC should also provide complete transparency. The reward mechanism and the freedom of asset transactions, and decentralization is a good way to solve these problems and can provide additional value. We are optimistic about teams with web2 creator resources and introduce these creators with transparency and higher returns. Come to web3 UGC platform.

Game UA

A summary of game UA projects , all centered on establishing user portraits, combined with the three dimensions of on-chain, off-chain and social, are divided into customer acquisition direction (Carv) and operational strategy direction (Helika), but all customer acquisition platforms often face retention problems Challenges, but it is undeniable that player data is valuable, and this value will increase with the growth of the number of users. If the application is optimistic about bringing large-scale adoption, game data analysis can capture part of the value.

Extended to the game distribution platform, the distribution platform with infrastructure and tools as its core products in the last cycle gradually lost its competitiveness. Now it can determine whether the distribution business can be done. To get up, we still need to rely on core game products that can bring in a large number of users.

User/Fan engagement

The project mainly involves entertainment, sports events and film and television related industries , according to the cooperative relationship with the IP party, it can be roughly divided into two forms: direct operation by the IP party and IP authorized cooperative operation. The IP cooperation form puts less pressure on the platform operation, but its effect depends on how many resources are given by the IP party, while the IP party directly operates It will focus more on operations, but it can usually better connect IP events/content with terminal products, and at the same time grant users more ancillary rights through NFT, and the performance of fan incentives and feedback may be better. In addition, through tracking and observation of several projects, we found that platforms with a certain community base (such as Karate Combat) can turn existing users towards IP fans more often than a fan engagement that starts from scratch and is completely based on IP. Be easy to succeed. Next, we will focus on users/audiences/fan groups with high IP value, overlap with game/betting portraits, and more direct cooperation IP targets.

Institutional Service

The institutional service track can be divided into the following sub-tracks:

Trading/brokerage services: including exchanges, liquidity providers, brokers/dealers, clearing/settlement, etc.

Asset management: including fund management, high-frequency trading, arbitrage, custody, etc.

Bank/payment: including payment processors/deposits and withdrawals, card issuance, bank-related services, etc.

Other services: including trading technology providers, etc.

The overall trends in this track include:

The institutional service track is expected to be in the next few years will continue to maintain steady growth.

Compliance is an important trend in the institutional service field, and every company is actively carrying out compliance construction.

The division of responsibilities among various service providers is becoming more and more clear. Each participant in the track focuses on their core responsibilities. Mutual checks, balances and supervision in various fields are conducive to making the market operate more honestly and efficiently.

The market share of PB service companies is expected to gradually increase. The fields they provide are still blank in crypto but are already relatively mature companies in traditional fields ( ECNs, Fully regulated clearing house, Cross Margin capabilities) deserve continued attention.

During this window period for ETF application, crypto-native service companies will be impacted by traditional financial companies, and the demand for compliance-grade products will further increase. The market structure will be reshaped.

Europe is a hot area for the development of institutional services, and emerging markets such as South America also have certain potential.

Bitcoin

Although Bitcoin has attracted much attention recently due to inscriptions, etc., the Bitcoin chain There is no global shared state, and the entire idea (state, account, calculation model) of Ethereum is very different. Therefore, in the medium and long term, the construction of Bitcoin infrastructure and applications needs to adopt different ideas.

On this basis, it is worth paying attention to Taproot Assets, Rollup, Lightning Network, etc.; in addition, new technology routes such as Statechain are also worth tracking.

Sidechain

Sidechain technology including Stacks The route has long occupied the mainstream narrative of Bitcoin's second-layer network, because the technical baggage is relatively small (it can be directly outside the chain), it can achieve relatively high programmability, and it is easy to achieve better ecological effects. But it will mainly rely on technologies such as cross-chain and anchoring with the main chain, and may be attracted more traffic and attention by other new technical routes.

Layer2

Many of the current so-called Bitcoin Layer2 are from In terms of core technical principles, it is still more like a side chain form, but it will build a complete project framework of execution, settlement, verification/challenge, and DA according to the technical model of Ethereum. The difference between the current Bitcoin Layer2 projects is mainly due to the different technology stacks selected at each level, such as Cosmos SDK, OP Stack, Polygon zkEVM, Taiko, etc. used in the execution layer; the DA layer is currently implemented by some third-party projects, and there are also Implement it yourself; in addition, a layer of "account abstraction" or the integration of multi-chain wallets is usually done to support both the address formats of Ethereum and Bitcoin to facilitate user operations.

Client-side validation

Includes RGB and Taproot Assets in The client-side verification technology within it can realize functions such as asset issuance and trading while minimizing the footprint on the chain, which deserves continued attention.

Lightning Network

Lightning Labs ) plans to launch stablecoins and other assets on Taproot Assets next year. In addition, the promotion of products that can achieve native asset income, including LSP (Liquidity Service Provider), is also worth looking forward to.

BRC20 class

BRC20 class assets are heavily dependent on For some specific infrastructure, such as indexer, etc., it is worth paying attention to this part of the infrastructure and some new asset types such as ARC20, etc., but it is necessary to pay attention to the technical implementation risks, etc.

DLC

Although DLC was proposed earlier, Because there was less demand before, promotion was more difficult. With the subsequent large-scale expansion of ecological construction, the application of DLC technology may become more popular, especially after cooperation with some oracles. However, attention should be paid to some centralization risks introduced during the implementation of DLC.

DePIN

DePIN is an easy buy in a bull market Quantity track. Similar to games, DePIN is also a track where it is easy to break out of the circle and convert traditional users, so it has also attracted industry attention. DePIN has several key elements: 1) It is decentralization, as well as the gameplay and mechanism. It can be said that it is the lifeblood of DePIN. Investment in DePIN projects needs to look at the mechanism first. 2) It’s timing. A good mechanism needs to be matched with good timing. It is definitely easier for projects launched in the early stages of the bull market to acquire customers early, so the team is required to be sensitive to the web3 market. 3) It is the fundamentals of the industry, the type of hardware selected by the project party, and the user characteristics targeted at it, which may determine the success or failure of the project. The following is a classification based on hardware type:

Pay attention to toC scenarios and markets where similar hardware has not yet become popular, revolutionary changes may occur:For hardware that is frequently used in ToC (such as wearable devices), the gameplay and mechanism of Web3 actually provide project parties with a better and more efficient product crowdfunding channel, reducing barriers for both users and merchants. In the world of DePIN, because there are clear token incentives, users will be more motivated to purchase hardware (the money will be paid back quickly), and merchants can even pre-sell before producing; with flexible cash flow, merchants can purchase later. Do fundamentals, such as how to enrich the software ecosystem, how to link it with other hardware, how to empower tokens in the ecosystem, etc. Especially for underdeveloped areas, some hardware may never be purchased without DePIN, but the speculative mentality of early participation in mining may lead to large-scale real hardware popularization.

Be careful about improvement hardware: For ToC placement type or low-frequency but necessary hardware (such as routers), the user group and With a large retention volume, DePIN can be an opportunity to improve the experience. In theory, DePIN realizes the redistribution of costs and income by redistributing resources and needs between suppliers and users, forming more reasonable unit economic benefits and making user-side services cheaper. But there are also difficulties: 1) Whether it is technically possible to achieve decentralization over centralized solutions. For example, many decentralized computing or storage solutions are even more expensive and less efficient than centralized solutions; 2) It is commercial Will it offend against the interests of large centralized manufacturers? The rigid demand and large holdings mean that there are many large manufacturers in this direction. They have accumulated many years of users, brands, and funds. If the DePIN solution cannot make a revolutionary breakthrough in fundamentals , or cannot unite to surpass the influence of big brands, it is actually difficult to compete with web2 competitors.

Watch the direction of unique mining machines: For other low-frequency uses, hardware that is dispensable in daily life, even for mining DePIN may bring short-term returns for specially purchased directional mining machines, but it may not necessarily create user stickiness. It is not ruled out that DePIN has really cultivated new user habits, but there is a certain degree of chance and it is difficult to predict. We will analyze specific projects in detail.

< section data-v-42644c81="">