In the infrastructure sector with "overcapacity", who is making money? | Rhythm Finance

Infrastructure seems to be an eternal topic for Crypto.

Although Infra is not as directly face-to-face with users as DeFi and other applications, it can always get the support of top institutions behind the scenes. From Crypto Native to Old Money, starting a business with infrastructure as a support seems to have become a "stable" business. The track is stable, the model is identifiable, and you can collect rent while lying down, and you can make a profit without losing money. Believers in technology and resource tycoons do not seem to want to get involved in high-risk Mass Adoption. More and more high FDV projects are emerging like factory copy and paste.

In the end, the speculators' 100-fold dream and the valuation bubble of the project were ruthlessly punctured by the market. VC coins became a street rat that everyone shouted and beat overnight, and endless infrastructure brought the ultimate question to the industry: Do we really need so much infrastructure? With this question, this article selects mainstream projects in the field of Crypto infrastructure as representatives to analyze their operating conditions, giving readers a perspective to judge whether the current market infrastructure should be cleared.

BlockBeats Note: There is no clear regulation in the Crypto market, so unlike traditional listed companies, the project party is not obliged to fully disclose its specific performance. Among them, the project party's income sources include wallet transfers, off-chain income, and token sales. Therefore, this article will use the publicly disclosed on-chain data that can be collected as the main reference and evaluation basis.

Oracle

Oracles can not only be used in Crypto to provide quotes for various dapps and fill the information gap between on-chain and off-chain, but also serve traditional financial products such as bonds and credit default swaps. At the same time, in Crypto where regulatory compliance has not yet been fully followed up, we cannot obtain the complete revenue situation after third-party audits through existing channels, and only obtain data through public channels on the chain for analysis.

In the oracle track, TVS (Total Value Secured) is usually used as a key measurement indicator. According to DefiLlama data, Chanlink still maintains its kingly style, with a market share of 44% in the total TVS, which is monopolistic. The TVS market share of the other top 5 protocols is WINKlink (15%), Chronicle (13.59%), Pyth (8.95%), and Swithboard (3.88%). However, WNLink and Chronicle are mainly used in the TRON ecosystem and MakerDAO respectively, and both rely on their own "big trees" for shade and do not have widespread market applications. Therefore, Chainlink and Pyth were selected as the analysis objects in the oracle field.

Oracle TVS share, data source: DefiLlama

Chainlink

As a veteran oracle project, Chainlink's main businesses include data feeds, CCIP, VRF, and contract automation (keepers). Its market share has remained at this level since it broke through 40% in 2020, and it has maintained an absolute monopoly in the market. Its business not only covers on-chain, but also includes cooperation with off-chain entities, such as SWIFT and Google Cloud. Chainlink's market application scope and scale are currently unmatched by other oracle protocols or even other infrastructure projects. In terms of specific business share, AAVE V3 accounts for 51% of Chainlink TVS, the largest share and a total TVL of 11.4 billion. The last four are Spark, Venus Core Pool, Compound V3 and Eigenpie.

Chainlink TVL distribution, data source: DefiLlama

According to Dune data, since entering the bear market cycle in September 2022, Chainlink's monthly revenue has remained at around US$4 million. In this cycle, Chainlink's overall on-chain revenue has not returned to the level of the 2021 bull market, barely reaching US$4.7 million at the peak of the small bull market in March this year.

It is worth noting that AAVE, as Chainlink's largest revenue contributor, has repeatedly set new highs in protocol fee revenue this year, and its overall performance is close to the level of the previous bull market. At the same time, since the end of 2023, AAVE's daily active users have continued to grow, and it is not affected even when the market cools down. However, Chainlink's revenue has not been significantly improved as a result.

In terms of revenue share, Ethereum was the only one in the last cycle, contributing more than 60% of revenue. Since the rise of various L2s, Ethereum's phenomenon of being too big to fail in Chainlink has been alleviated, and its share has gradually declined. The share of emerging L2s represented by Optimism and Arbtrium has increased significantly. On the one hand, Chainlink's revenue sources are more dispersed and diversified, but on the other hand, it also reflects the split of ecological liquidity brought about by L2.

Chainlink revenue sources and proportions, data source: Dune

CCIP, as Chainlink's flagship product for entering the cross-chain track, was officially released in July 23. Chainlink supports almost all mainstream public chains and protocols on the market, and at the same time has first-hand data information. CCIP's business integration can be said to be "time, place, and people".

According to Dune data, CCIP uses Cross-chain fees as its main source of business income, and is mainly contributed by mainstream L2s such as Arbitrum, Base, and Optimism. Cross-chain income depends largely on the market's prosperity and is sensitive to market reactions. In March of this year, its cross-chain income peaked at about US$258,000 under the favorable catalysis of the approval of BTC ETF. After the on-chain market cooled, Q2 revenue decreased by 30%, and reached an extreme freezing point in the Q3 market, with revenue falling again by 70% and finally maintaining at $25,000 per month.

Cross-chain fees and sources, data source: Dune

In terms of token performance, its native token $LINK did not have any value capture function before the update, and with the continuous release of node rewards, the micro level will only continue to dilute the price of $LINK. Since the release of Token Economics 2.0 in 2022, the practicality of its tokens has changed, mainly through the staking function of $LINK. In Staking1.0, only the community and node operators are allowed to stake tokens. After the release of version 2.0, not only the scale of staking has been expanded, but also the threshold for staking has been lowered and the unstaking mechanism has been opened. Currently, the circulation of $LINK is about 60%. Although it has not been fully unlocked, the token price of $LINK has performed well in this bull market thanks to its staking mechanism.

Pyth

As a rising star, Pyth Network adopts a "pull" price feeding service that is different from traditional oracles, improving the previous push-type service to on-demand access. DeFi protocols can enjoy Pyth's price feeding service by developing corresponding contracts, saving more time and development costs to polish their own products. With its simple service model and the support of the Solana ecosystem, Pyth has become Solana's largest oracle protocol.

According to DefiLlama data, Pyth's business is concentrated in Kamino Lend, with a TVL of US$1.32 billion, accounting for about 30% of the total TVS. At the same time, Pyth's total TVS is increasing over time. This year's total TVS is more than double the high point in 2022, and its development scale is still expanding.

Pyth TVL(TVS) growth, data source: DefiLlama

Pyth TVL(TVS) distribution, data source: DefiLlama

There is no public information on the market to query Pyth’s income and expenses, so we use the operating data to "see the leopard in the tube" to infer whether Pyth is in good overall condition. Compared with the leader Chainlink, the "dark horse" Pyth has an absolute advantage in the scale of Price Feeds on Solana. On a weekly basis, Chainlink's number of price feed transactions on Solana is far less than Pyth, with an average of less than 1 million transactions per week. Pyth basically maintains above 1 million transactions, accounting for 85% of Solana's total price feed transaction volume. As Pyth's "happy hometown", Solana is its rear base, and it is reasonable. Considering the business scale of the EVM chain, Pyth is in a period of steady growth. After the average weekly transaction number exceeded 100K, it gradually stabilized and mainly used public chains such as Arbitrum, Fantom, and Optimism.

Left: Number of Solana transactions processed by Pyth (compared with Chainlink); Right: Number of transactions processed by Pyth EVM chain and source, data source: Dune

However, Pyth still has a long way to go in the market expansion of the EVM chain. According to Dune data, Chainlink still has an absolute monopoly on a weekly basis, accounting for about 92.7% of the EVMs market share. Although Pyth is developing rapidly, the number of transactions processed by its EVM chain is only a fraction of Chainlink.

Comparison of the number of transactions processed by Pyth and Chainlink EVM chains, data source: Dune

It is worth noting that according to Dune's data, for the entire Crypto market, the largest number of transactions processed by Chainlink comes from Solana, and the number of transactions from EVM accounts for about half of Solana.

Number of weekly transactions processed by Chainlink (Solana VS EVM), data source: Dune

In the context of the oracle track gradually becoming a red ocean market in Crypto, whether from the cost side or from the ecological side, Pyth needs to solve the "WHY YOU" problem for users in order to grab more EVM shares.

From the perspective of token investment, $PYTH's value capture is similar to $LINK, and community voting governance can be carried out through staking. In addition to the staking function, users cannot share the dividends of Pyth's development. At the same time, the total number of $PYTH is 10 billion, and the current circulation is only 36%. After the large-scale unlocking in May and the cold market, the price of $PYTH has been hit hard. It has fallen all the way from the highest of 1.2 to the recent 0.3.

Although Pyth's fundamentals are in full swing, investors in its tokens need to understand whether these businesses are related to the tokens they hold.

Cross-chain interoperability

Since Crypto has formed a multi-chain pattern with Ethereum as the leader and other public chains as the stars, cross-chain has made up for the liquidity dispersion problem caused by the fragmentation of public chains and has become an indispensable part of crypto infrastructure.

In the field of cross-chain interoperability, the three major cross-chain protocols Wormhole, Layerzero and Axeler and cross-chain applications Across and Stargate were selected. This article only analyzes and compares the profitability of cross-chain infrastructure from a financial and operational perspective. However, for cross-chain users, "there are tens of millions of public chains, and safety is the first", and the security of user assets should always be the first priority in project evaluation.

Wormhole

Wormhole is a cross-chain communication protocol developed by Solana and Certus.One and launched on September 22, 2021. It was initially used to realize cross-chain token assets between Ethereum and Solana. As of now, Wormhole can support asset transfer functions on multiple chains including BSC, Avalanche, Terra, Base, etc. Wormhole, as the one with the highest market value among the three major chains, has a slightly different fee structure. According to official disclosures, for Wormhole Gateway, the GAS of the source chain is currently the only cost, and users only need to pay GAS.

According to Dune data, as of now, the vast majority of cross-chain transactions still come from Ethereum to Solana, accounting for about 50%, followed by Bsc, accounting for about 15%. Unlike the market entry point of other cross-chain protocols, the L2 share of Wormhole cross-chain accounts for a small proportion.

Wormhole cross-chain sources and number of transactions (from Ethereum to other chains), data source: Dune

Due to the design of the protocol fee mechanism, we cannot obtain intuitive revenue data. Therefore, we measure Wormhole's business performance from the perspective of user activity and transaction volume. Wormhole performed well in Q1 this year, with the highest daily transaction volume approaching 50,000. The increase in transaction volume in January was mainly due to the active on-chain transactions caused by the wealth-making effect of Solana's ecosystem MEME Coin, and cross-chain transactions to Solana accounted for half of the total, which eventually led to a surge in transaction volume in January. However, as the Solana MEME narrative gradually weakened, Wormhole's transaction volume also declined visibly. For most of this year, it remained stable at around 3,000 to 5,000.

Wormhole daily cross-chain transaction volume, data source: Artemis

For token investors, Wormhole's native token $W has limited value capture capabilities. The main function of $W is to conduct on-chain governance after staking. Unlike other cross-chain protocols, Wormhole uses the POA consensus mechanism, where validators are composed of 19 "trustworthy" large institutional validators and do not need to stake tokens when verifying inter-chain messages. This not only reduces the application scenarios of the token $W, but also the validators under the POA mechanism lack sufficient economic incentives, and the cost of doing evil is significantly reduced.

LayerZero

As the underlying protocol for full-chain communication, LayerZero's current main application scenario is the transfer of token assets, as well as NFT, cross-chain governance and other operational use cases. LayerZero's revenue mainly comes from cross-chain information fees: when users transfer their token assets from the source chain to the target chain, the LayerZero protocol will charge fees from the source chain in the form of currency. In the most recent quarter (2024 Q3), LayerZero's main sources of income were Arbitrum, BNB Chain and Ethereum. At the same time, in Q2 of this year, due to the reduction of GAS, the proportion of cross-chain transactions of Arbitrum and Optimism networks increased significantly. As the head protocol in the cross-chain track, its integrated dapps also successfully broke through the 50,000 mark this year.

Left: LayerZero's revenue source ratio; Right: LayerZero's cumulative number of integrated applications, data source: Dune, LayerZero official website

According to Dune data, LayerZero's revenue in the three quarters of 2024 was: US$37.6879 million, US$12.8626 million and US$972,700 (as of August 31). It can be clearly seen that LayerZero's protocol revenue and began to decline sharply after entering Q2 and never recovered. On the one hand, because LayerZero officially issued tokens in June and took anti-witch measures before that, many data whitewashings were cleaned up; on the other hand, the overall market environment was bearish in May, and the sharp decline of major public chain tokens made its revenue worse.

LayerZero monthly revenue and share, data source: Dune

LayerZero daily transaction statistics, data source: Dune

LayerZero's overall operating conditions are closely related to the activity of the on-chain market. When the on-chain market is bullish, Layerzero's operating data also takes off. After May of this year, the market went bearish, and its revenue was close to "foot-cut", slightly higher than the level of the previous round of bear market. This is not only due to the market environment, but also the substantial reduction in GAS fees brought about by the Cancun upgrade is also an important factor in the decline of Layerzero's revenue by an order of magnitude.

LayerZero's native token is $ZRO, which had a market value of 1 billion when it was just launched in June this year, but has now been halved to 433 million. $ZRO is Layerzero's governance token and can be consumed during cross-chain messaging. In terms of fundamentals, Layerzero's P/F ratio is still overvalued because the decline in market value is smaller than the decline in revenue.

Axelar

Axelar's business model still uses the traditional "rent collection" method. Transactions and smart contract executions on the Axelar network will incur corresponding fees, which are paid in AXL tokens and bring direct income to the network. Users who conduct cross-chain transactions will pay GAS to the Axelar network, which will then redistribute it to stakers. Axelar's main expenditure is the $AXL token rewards paid to validators. At the same time, Axelar is also developing AVM, intending to open up cross-chain smart contract development. According to official data, the main cross-chain transactions currently come from BNB Chain and Sei. The transaction volume of the two is US$3.4 billion, accounting for about 40% of the total transaction volume.

Axelar official statistics on cross-chain transaction distribution map, source: Axelar official website

From 2024Q1 to 2024Q3, whether in terms of transaction volume or cross-chain fees, Axelar's various indicators denominated in US dollars are not optimistic, with a sharp drop of more than 50% year-on-year, showing an accelerating downward trend. From the perspective of price-earnings ratio, Axelar's P/F ratio in the third quarter did not decrease but increased. The main reason is that its market value has declined by about 30%, which is less than the halving of cross-chain fee revenue, so the P/F ratio is generally inflated.

Left: Changes in Axelar's quarterly financial indicators; Right: Axler's market capitalization and transaction fee comparison chart, data source: Tokenterminal

In terms of user data, Axelar's daily active users formed a double peak in mid-January and mid-July this year, with the highest daily active users reaching 4.5K. But after mid-July, it gradually fell back to around 1.5K. Compared with 2023, Axelar's users this year remain active as a whole. It is worth noting that the proportion of GMP (General Messaging) in the Axlear network has continued to increase, and has almost completely replaced simple asset cross-chain, accounting for 98% of the total active users.

Axelar monthly activity data, data source: Axelar official website

Axelar has currently linked 66 chains including Ethereum, BNB Chain, Arbitrum, Fantom and other mainstream public chains. As the number of linked networks increases, the threshold for validators will also increase, and the corresponding token rewards will also rise. Initially, such a plan would lead to an increase in the inflation rate of $AXL. After the official realized the problem, it replaced it with a GAS burning mechanism. The change in the overall token inflation mechanism is beneficial to the $AXL token investors, and its annual inflation of 11.5% can be offset by 10% through GAS burning.

Across

Across is a cross-chain protocol supported by UMA, which relies on UMA's optimistic oracle mechanism to achieve cross-chain between L2 and L1. Currently supported chains include mainstream L2 public chains such as Arbitrum, Optimim, Base, Linea, and Blast.

In terms of the technical design principle of cross-chain, Across introduced the relayer ReLayer, and relied on it to lock the token assets of the source chain and release the corresponding tokens in the target chain. After the relay execution is completed, the optimistic oracle will perform subsequent data verification to ensure asset security and reward the relayer. The Relayer in the Across network relies on the liquidity of the corresponding tokens provided by the LP on the target chain when bridging. If the liquidity is exhausted, the user's cross-chain cost will surge.

Therefore, the good operation of Across is inseparable from the participation of the left and right hands "Relayer" and "LP". Both parties bear all the risks involved in any bridge event on Across, and users pay fees to them when using cross-chain services. On the one hand, the operating model of the liquidity pool is essentially similar to AAVE or Compound, and its pricing scheme also refers to the pricing design of AAVE and adopts the interest rate pricing model: the interest rate R is calculated through the model, and the LP fee is calculated based on a one-week trading cycle. On the other hand, the relayer will advance GAS when crossing the chain, and will receive fee compensation after the transaction is successfully verified. Therefore, when evaluating whether Across operates well, transaction volume and transaction fees will be more fair evaluation criteria.

According to DeFiLlama data, Across's transaction volume in the past month was $637.94m, ranking sixth among cross-chain applications. As of September 11, Across's cumulative users have exceeded 2 million, and monthly active users have stabilized at around 200,000.

Cross-chain application ranking (sorted by transaction volume), data source: DefiLlama

At the same time, according to data provided by Dune, the cross-chain transaction fees caused by Across in Q2 this year are already comparable to the daily transaction fees at the end of the last cycle. It has reached the level of 60,000 US dollars per day. Although Across is not as famous as other cross-chain applications, it still has a certain competitiveness in the market.

Across daily cross-chain transaction fees and composition, data source: Dune

In terms of cross-chain transaction volume, Across's cross-chain volume reached a high level of 2.4 billion US dollars in Q1 at the beginning of the year. Subsequently, the MEME craze in Q2 maintained the cross-chain transaction volume above 2.5 billion US dollars. However, contrary to this, although the number of on-chain transfers has increased, the handling fee has not increased to its due level. Instead, it reached its peak in Q1 this year, then fell 33% in Q2, and halved 66% in Q3. The reason for the decline in Q2 is that the Cancun upgrade has caused a significant drop in L2's cross-chain GAS. In Q3, the transaction volume declined due to the sluggish market.

Across's financial indicators, data source: TokenTerminal

From the perspective of token investment, Across's native token $ACX has weak value capture capabilities. $ACX can be used to manage DAO funding decisions and as a lock-up reward for LPs. Across has proposed an LP lock-up reward plan. The longer the lock-up time, the more multiplier rewards and $ACX token incentives you will get. Since the Snapshot proposal was passed in September 2023, Across has increased the APY % of the liquidity pool by 50% and reduced the original 3x multiplier to 2x.

Obviously, for the Across project itself, there is no direct correlation between its profitability and the price and market value of the token. Although the transaction fees of the Across network have been cut in half and cut in half again, the market value of its token $ACX remains stable, maintaining at around US$120 million. Therefore, the decrease in transaction fees has led to an increase in the P/F ratio, and the overall project is overvalued. But for investors, the overall high P/F ratio does not seem to have a direct impact on the investment decision of $ACX.

Across transaction fees and circulating market value comparison chart, data source: Tokenterminal

Stargate

Stargate is officially launched by LayerZero and is a composable liquidity transmission protocol developed based on LayerZero technology. Stargate is positioned to become a cross-chain application for the entire chain DeFi. It currently covers more than 10 public chains including Abriturm, Scroll, Base, Linea, etc. According to DeFiLlama data, its scale has reached $1b in the past month, second only to Arbitrum Bridge, ranking second.

Data source: DefiLlama

In addition to cross-chain interaction, Stargate also has businesses such as building liquidity pools and staking mining. The Pool model is similar to Across. When building a full-chain cross-chain, LP is required to provide corresponding liquidity on the target chain of the Stargate cross-chain. Investors can provide liquidity as LPs and obtain corresponding returns. On the other hand, users can stake LP tokens to obtain $STG token rewards.

Stargate's revenue channels mainly come from the handling fees charged for token transfers. For non-$STG token transfers, a fee of 0.06% will be charged for each transaction. Among them, LP and veSTG holders will receive a fee allocation of 0.01% respectively.

However, Stargate's transaction fees have been declining since the beginning of the year. The highest reached $3.34 million, and it was as low as $196K in Q3. However, compared with Across, Stargate's business flow is still huge, and the transaction volume remained between $3 billion and $5 billion in the first two quarters, but it declined severely in Q3, only $226.82M.

Changes in Stargate's various financial indicators, data source: Tokenterminal

The protocol fees collected by Stargate began to plummet after May this year, and the total fee income in Q3 this year evaporated 90% compared with Q2. However, the circulating market value of $STG did not decrease with the reduction of total fees, and it has always remained at around 120 million. As a result, the P/F ratio surged in Q3, and from this perspective, the price of $STG is seriously overvalued.

Stargate transaction fees and circulation market value comparison chart, data source: Tokenterminal

Compared to $ACX, Stargate's native token $STG adopts the ve model design in token economics, which has stronger value capture capabilities. Token holders obtain veSTG by staking $STG, vote on Stargate's fee standards, liquidity pool adjustments and other matters, and participate in community governance. At the same time, as the time of staking STG tokens is extended, users can obtain additional veSTG rewards to increase their weight in community governance, and the positive flywheel can be accelerated. On the other hand, users can provide $STG to the platform to increase liquidity, or they can obtain $STG rewards through mining. Overall, the market value of $STG is relatively stable.

Storage

If Web3 can achieve true Massive Adpotion in the near future, decentralized storage of data is crucial and will become an indispensable part. Although the overall data storage scale of Web3 is insignificant compared to traditional Web2 storage, this also means that future Web3 storage projects have a higher ceiling and a huge stock market. In the field of storage infrastructure, we selected two major projects, Filecoin and Arweave, for comparative analysis.

Filecoin

Filecoin adopts a pay-per-use model, rather than Arweave's one-time deal. Filecoin's storage providers (i.e. miners) mine $FIL by providing storage services to Filoin users or committing storage capacity on the network. To encourage long-term network coordination, 25% of the block rewards received by miners can be used immediately, while 75% will be unlocked linearly within 180 days. Secondly, if users need to use storage services, they need to pay a certain $FIL fee. Similarly, when they need to retrieve and retrieve in the future, users also need to pay a certain fee.

In general, Filcoin's operating income is mainly composed of user interaction GAS and fines and confiscations from miners, and the $FIL token reward to miners is the main expenditure under its business model. At the same time, Filecoin is actively transforming to on-chain computing and has developed FVM to facilitate developers to develop applications on Filecoin to complete ecological construction.

According to Messari data, the number of Filecoin users with storage scale greater than 1,000 TiBs is increasing. Such users are usually enterprise-level users with large data storage and good stickiness.

Changes in Filecoin users with storage greater than 1,000TiBs, data source: Messari

According to TokenTerminal data, Filecoin's fee income has continued to increase from the beginning of the year to the end of June, with an increase of up to 151%, but the overall fee has dropped sharply since entering Q3. From the expenditure side, Token incentive, as a block reward, has begun to decline rapidly since the end of Q1. This situation is not only caused by the spiral design of Filecoin's token economics, but the main reason behind it is the increasing loss of miners. According to Filscan data, since the end of 2023, Filecoin's active storage providers have been "running away". From a traditional valuation perspective, the P/F Ratio was artificially high in Q3 due to the reduction in fees.

Changes in Filecoin's various financial indicators, data source: Tokenterminal

Statistics of active storage providers, data source: Filecoin official website

In terms of token value capture, its token $FIL is not only used as GAS in the Filecoin network, but can also be used as a token incentive for "miners" to provide storage services, and "miners" need to pledge $FIL tokens when providing storage services to ensure the security and service quality of the network. Filecoin's overall token economics design presents a spiraling pattern.

Filecoin’s price in the last bull market was very exaggerated. At its peak, the price of $FIL reached 200 US dollars per coin, and its FDV reached an exaggerated 380 billion. The reason for the price surge is that the unlocking amount of $FIL is small. On the other hand, since the miner reward is based on $FIL, when the price of the coin rises, it will attract miners to enter the market, and miners will buy more $FIL for staking. Thanks to the design of token economics, the price of $FIL finally soared.

However, after the Bitcoin ETF ushered in the bull market this year, the price of $FIL did not usher in a surge like the previous round of bull market. As of the time of writing, the price of $FIL has increased by only 4.6% in the past year. In contrast, the price of Arweave's native token $AR has increased by as much as 313% in one year.

Data source: Coingecko

On the one hand, the market environment is not as expected; on the other hand, unlike $AR, which is close to full circulation, the liquidity of $FIL is about 30%, and there are a large number of unlocked tokens, which puts a certain amount of selling pressure on the overall token price. Investors need to carefully consider between investment projects and project tokens.

Arweave

Arweave is different from Filecoin mentioned above. Its selling point is to realize decentralized permanent storage of user data on the chain. The Arweave protocol is built on the HTTP protocol. Users can access Arweave through Web2 browsers such as Google Chrome and allow individuals with idle hard disk space to store data in exchange for AR tokens

Arweave adopts a one-time fee business model. Users only need to pay the corresponding fees according to the capacity of the stored data to achieve permanent storage. Part of the one-time fee paid by the user is used for the initial storage cost, and the other part goes into the donation Arweave fund for future storage costs.

Although Arweave's user uploaded data is the same as the previous bull market, or even slightly surpassed, reaching a maximum of 8.79T/month. However, compared with Filecoin's user uploaded data, it is still not as expected. According to Starboard data, Filecoin's daily upload volume has reached 1.5E (where 1 EiB=1024 TiB), and Arweave's storage scale is an order of magnitude different from Filecoin.

Arweave daily upload volume, data source: ViewBlock

Filecoin daily upload volume, data source: Starboard

According to the monthly estimate, taking Arweave's average storage of 4T as an example, according to the calculation results of the ar-fee website, it costs about $70,000 to store 4T of data. According to CoinGecko data, the data storage fee for Filecoin to store 4T is $0.76 per month. With the same cost of $70,000, Filecoin can be used for storage for 7,675 years. 7,675 years is obviously not worth mentioning compared to "permanent storage", but for a company or individual, this is almost equivalent to eternity.

Arweave storage fee calculation, data source: https://ar-fees.arweave.dev/

When we re-examine Arweave's "pay once, store forever" business model, such a narrative is not a sexy story from a fundamental logic point of view. First of all, the essence of Arweave is similar to a "time capsule", and its positioning is aimed at small and beautiful projects. It is not attractive to large enterprises. The data that large enterprises need to store is usually private information rather than public information; on the other hand, from a cost perspective, many of the company's data does not need to spend such expensive storage fees for permanent storage, and it is more cost-effective to adopt Filecoin's storage solution. Therefore, in terms of fundamentals, it is difficult for Arweave to attract large customers, which is why its data upload scale remains at 4T.

Although Arweave has a limited audience and a relatively low product ceiling compared to Filecoin, from a micro-data perspective, Arweave's transaction fees are only slightly lower than the previous bull market if priced in currency. The transaction fee in September 2021 was 4468 $AR, worth about $223,400 (50*$4468); the transaction fee in August 2024 was 6685 $AR, worth about $167,125 (25*$6685).

Arweave currency-based transaction fees, data source: ViewBlock

For investors in $AR tokens, $AR assumes the basic payment function of currency. Arweave's native token is $AR. On the one hand, $AR can be used as a means of payment to pay for data storage. When users store data on the Arweave network, they need to use AR tokens to pay a one-time storage fee. These fees ensure the permanent storage of data. On the other hand, miners receive AR tokens as rewards by storing and verifying data. For the token price performance, please refer to the previous analysis.

DA Data Availability

The large public chains we are familiar with usually adopt an integrated, monolithic design, with a single blockchain processing all transactions. However, whether it is social division of labor or Crypto, the development of refined division of labor seems to be inevitable. Since the development of modularization, data availability (DA) has become an indispensable component of Ethereum Lego blocks. Among them, Celestia, EigenDA and Avail are more representative. Due to the limited data of EigenDA and Avail, we selected Celestia for analysis.

Celestia

Celestia's native token is $TIA. Celestia can support developers to develop new public chains and use $TIA to pay GAS. Users can also stake $TIA to get rewards. Due to the popularity of $TIA airdrops, the benefit effect was obvious at the beginning of its launch, up to $20/TIA. After the market cooled, the price of $TIA gradually returned to normal levels.

Celestia's user profile is mainly concentrated in Rollup. L2s on the market use Ethereum for data availability and transaction settlement, which incurs high costs, while the use of Celestia will cause a "qualitative" change in costs. Taking Arbitrum and Optimism as examples, if Celestia is chosen instead of Ethereum, then every 100 million transactions will incur DA costs of $32,598 and $15,333 respectively, reducing the total cost by 89% and 92% respectively.

Arbitrum and Optimism compare the DA cost savings of Celestia on Ethereum, data source: hashedem.substack.com

Modular projects like Celestia have brought great convenience to L2. Reflected in the fundamental data, Celestia's market share is gradually rising. According to Blockworks data, it has reached a market share of 44%. Although Celestia's security cannot be compared with Ethereum, the market needs low DA costs more than security. The low-price strategy has effectively helped Celestia expand its market scale.

DA market share (according to Data post statistics), data source: Blockworks

It can be seen intuitively from the figure below that after each project adopts Celestia, the average cost can be reduced to about 80%-90% of the original.

Cost savings of various projects adopting Celestia, data source: numia.data

Since Celestia can provide DA services at such a low price, is Celestia still making money?

According to Aremis data, Celestia's average daily uploaded data is about 3.5K/MiB, and the data upload fee denominated in USD has decreased from $0.5/MIB in June-July to $0.3/MIB in August-September, plus the GAS that needs to be consumed. So in fact, Celestia's potential annual revenue is relatively low, about a million dollars.

Although Celestia is still profitable at the data level, it has deviated from its valuation to a large extent. According to CoinGecko data, its FDV is 6.4 billion. Perhaps you don't have an intuitive feeling about FDV 6.4 billion, let's compare it to Chainlink at the beginning of the article, whose market value is currently about 7 billion and its FDV is about 12 billion. For Celestia's million-level revenue, it is obvious that the valuation of half of Chainlink is mixed with a lot of water. The actual business model and operation are not enough to support its inflated market value. Left: Celestia/Ethereum Data Post; Right: Celestia/Ethereum average cost per MiB, data source: Artemis

On the other hand, $TIA has only unlocked 26% so far, and will face a large unlock of $1 billion on October 30. At the same time, TIA does not have a maximum supply limit. And it is worth noting that Celestia Foundation announced on September 24 that it had completed a $100 million financing led by Bain Capital Crypto, Syncracy Capital, 1kx and other institutions. The official announcement of the completion of $100 million in financing before the unlocking is intriguing.

Token $TIA unlocking status, data source: DeFiLlama

For the DA track, although its main customers are various Roll ups, the current mainstream L2 still uses the Ethereum mainnet as DA, and very few L2s choose Celestia. In addition to whether Celestia's business model can be implemented, the more important thing is the delicate relationship between Ethereum and various L2s. Choosing Celestia is tantamount to openly singing against Ethereum. In this cycle where people can barely survive by huddling together for warmth, the market economy does not seem to be that important.

L2s

As the main theme of this round of infrastructure construction, L2 has achieved "lying down and making money" compared to L1, which often burns money crazily. With the technical support of OP STACK, the development cost of L2 has been significantly reduced, and "one-click chain issuance" has been truly realized. Usually, the cost of L2 is composed of sorters, DAs, development, etc. After the Cancun upgrade, the GAS of L2 has generally decreased by 50%-90%. While enjoying the advantages of L2's low GAS, the market has gradually discovered the shortcomings of L2. Do we really need so many L2s? With this question in mind, we divided them according to technical principles and market value, and selected Arbitrum, Optimism, Base from the OP system and Starknet and zkSync from the ZK system for analysis.

Comparison of transaction fees before and after the L2 Cancun upgrade, data source: hashedem.substack.com

Arbitrum

Specifically, Arbitrum's revenue experienced a sharp drop after Q1. Compared with Q1's $27.38 million, its Q2's $7.37 million and Q3's $3.34 million figures are unsightly, partly due to the sluggish market environment during Q2 and partly due to traffic competition from other L2s.

It is worth noting that due to the Cancun upgrade, Arbitrum's costs were also greatly reduced, so the net profit only fell by 7.3%. Compared with most other infrastructure projects that often see a 70% profit drawdown, Arbitrum's performance is far better than most of its peers.

Changes in Arbitrum's financial indicators, data source: Tokenterminal

As mentioned above, L2's biggest expense is paying DA and settlement costs to L1. Taking Arbitrum as an example, the L1 cost of processing 100 million transactions reached $25,000 per month. Of course, it would be lower if Celestia's service was used. If GAS covers these costs, then running L2 only needs to consider how to reduce the sequencer and initial development costs. When these costs are low enough, L2 is a "rent collection" business that makes money easily, but this is indeed the case.

Arbitrum revenue and expense statistics, data source: Tokenterminal

The user bottleneck has always been a shackle that restricts its development. Arbitrum's number of active users has been difficult to break through after reaching 10 million. Now, as the market cools, the number of monthly active users has also dropped to around 2.8 million. On the one hand, Arbitrum is likely to be overtaken by Base. On the other hand, Arbitrum's ecosystem is not only difficult to attract incremental users, but also needs ecological innovation to revitalize existing users.

Arbitrum monthly active user data, data source: Tokenterminal

Optimism

OP's overall operating conditions are similar to most of them. Its revenue peaked at $18.83 million in Q1, and was halved in Q2 and Q3, falling to $8.95 million and $3.32 million respectively. Although from the perspective of "political correctness", the Optimism project team has a close relationship with the Ethereum technical community and foundation, representing the successor of Ethereum's will in L2, and is more orthodox than Arbitrum's Ivy League circle. But the correct answer given by the market is not Optimism: Arbitrum is on par with it, and Base, which will be introduced later, is even better.

Changes in Optimism's various financial indicators, data source: Tokenterminal

It is worth noting that Optimism and Arbitrum have different retained earnings. Although its on-chain net income is increasing, Optimism is unable to make ends meet after considering the impact of token incentives. According to the U-standard pricing, it needs to pay $91 million, $34 million and 16M token fees in the first three quarters of this year, which is far higher than its operating costs.

Optimism on-chain net income, data source: Dune

In terms of active user data, although Optimism has only 1.6 million active users at most, which is not as beautiful as other OP-based L2s with 10 million, its lower limit is very stable, and the stable "decline" is around 1.2 million. Compared with its fellow sufferers in the ZK-based system, the overall user data of the OP-based system is much better, not only the ceiling is higher, but also the lower limit can be dragged down.

Optimism monthly active user data, data source: Tokenterminal

For Optimism, the overall atmosphere and positioning of the team are more tech-oriented. The OP Stack developed by OP Labs has helped many L2s to launch chains with one click, and has also been further upgraded to OP Succinct and ZK Stack to grab traffic. As for whether the vision of SuperChain can be realized, the answer needs to be given to the team and the market.

Base

Backed by the exchange CoinBase, Base was born with a silver spoon in its mouth from the beginning, and its fundamentals are much healthier than other L2s. As a rising star, it has now become a top player on the L2 chess table. In Q1 this year, Base's transaction fees soared 400% month-on-month. Prior to this, Base relied on "MEME" and "social" to rise to the sky. Due to its low transaction fees, Base has taken the old path of Solana and caused a MEME craze. BALD, BRETT, and DEGEN have successively ignited BASE's MEME fever. Subsequently, BASE integrated social applications friend.tech and Farcaster, becoming a veritable Social-L2.

After a set of combined punches, Base achieved a latecomer and surpassed Arbtrum and Optimism. Its Q1 revenue reached 27M, which is the same as Arbitrum. Despite the 70% decrease in Q3 revenue, its net profit was still as high as 5.84 million US dollars.

Base's financial indicators, data source: Tokenterminal

From the perspective of TVL, Base's success depends on two difficult-to-replicate social fissions. From the initial tepidity to the influx of a large number of new users and funds after the launch of Friend.tech, Base has been pushed out of the novice village, and its TVL has reached the level of 500 million. Subsequently, in March this year, Cancun upgraded GAS and Base launched the same MEME craze as Solana, and on-chain users began to become active. With the help of Farcaster and other Social-Fi, Base's TVL has reached the level of 1.5 billion. So far, Base's TVL is still increasing steadily.

Base TVL growth, data source: DeFiLlama

Unlike other L2s, user data is not the bottleneck restricting Base's development. Base has been steadily improving in terms of user data. The number of monthly active users reached 20 million in September, surpassing other "king"-level L2s. Relying on CoinBase's huge user traffic, the active users that are the most difficult to capture for other L2s have become Base's unique moat.

Base monthly active user data, data source: Tokenterminal

Similarly, Base is a sure win. In addition to the fact that the Base project team knows how to market and how to play, and maximizes traffic, Base's beautiful financial data is also due to the support of Coinbase. As the only sorter on Base, Coinbase has reduced the cost to the lowest level except for DA. Therefore, according to the data from TokenTerminal, except for a few points in time, Base's revenue far covers its costs most of the time.

Base income and expenditure, data source: Tokenterminal

Starknet

As the prince of ZK technology, Starknet has always been highly sought after, but its fundamentals are far from satisfactory. Since entering 2024, Starknet's revenue has only fallen and not rebounded, just like its currency price. Among them, the decline in 2024Q2 was the most serious, with only $147,900 in transaction fees compared to $9.37 million in Q1.

Q2 In addition to the negative impact of the depressed market environment, EIP-4844 and the Cancun upgrade are another important factor. Although the Cancun upgrade brings extremely low DA costs to L2s using Blob, it is a blessing in disguise, and it also brings a decrease in network transaction fees. Since the Cancun upgrade, Starknet's on-chain transaction fees have been sluggish. From the figure below, it can be seen intuitively that Starknet's revenue has been sluggish since Q1.

Starknet transaction fee quarterly changes, data source: Tokenterminal

From the user's perspective, Starknet has gradually become a "ghost town". In January 24, Starknet broke through the bottleneck of 1.2 million monthly active users, and then reached a peak of 2.2 million. Since then, active users have begun to lose in batches, and the user bubble was finally punctured. Today, the number of monthly active users is stable at around 130,000. Whether in level 1 or level 2, on Twitter or in group chats, Starknet’s presence gradually faded.

Starknet monthly active user data, data source: Tokenterminal

zkSync

The last of the four L2 kings to be listed, zkSync, has attracted much attention from the market and public opinion from its debut to the end of listing on Binance. zkSync released a test network in 2021, but the development progress was slow, and the main network was not officially launched until 2023. Although zkSync has a top capital platform, its ecological construction is not as expected. There is no phenomenal dapp that can even attract public attention, which is quite "thunderous but small". Before the issuance of its native token $ZK on June 17 this year, the main flow of the market was also focused on how to "brush data" for zkSync to win airdrops.

Changes in zkSync’s financial indicators, data source: Tokenterminal

Although zkSync’s revenue is as disastrous as most L2s, this only means that zkSync is making less money, at least it is still at the level of “profit withdrawal”. According to Tokenterminal data, its overall revenue can still cover its expenses. However, its net profit is at a lower level than other L2s, falling from $4.7 million in Q1 to $370,000 now.

zkSync income and expenditure, data source: Tokenterminal

It is worth noting that in July this year, zkSync suffered a double blow in revenue and the number of developers, but its active users rebounded, increasing from 2 million in June to 3.8 million, and then the monthly active users plummeted to 800,000. For zkSync, it not only needs to consider how to build its own ecosystem to attract funds and users, and strive to get bigger chips on the L2 battle table, but also how to avoid the vicious cycle caused by its own empty city.

zkSync monthly active user data, data source: Tokenterminal

Obviously, zkSync has realized the problems of weak ecology and loss of existing users in the fierce L2 competition. zkSync's TVL has doubled from 76 million to 140 million since September 20. The recent changes in TVL are mainly contributed by Venus Core Protocol, which increased by 2434% in one week. At the same time, zkSync has recently become the third L2 in terms of weekly active users, but there is still an order of magnitude difference with Arbitrum and Base. Although zkSync has been catching up recently, it still has a long way to go.

Left: zkSync TVL growth; Right: zkSync TVL share distribution, data source: DefiLlama

L2 weekly address activity ranking, data source: L2BEAT

As for the token economics of the above L2, except for Base, the native tokens of the remaining L2s almost only play a governance function. Although L2 is the right-hand man of Ethereum, users do not need to consume its native tokens in actual transactions, but still use ETH as GAS payment. This also determines that if each L2 does not develop an ecosystem that attracts users, it can only serve as a "transit chain" that saves money. Vitalik, who has the highest voice in the Ethereum ecosystem, recently published an article advocating for the alignment of the entire Ethereum ecosystem. Now, in addition to considering the objective economic laws of market operation, the development of the L2 track has added many uncontrollable "political" factors.

Summary

For infrastructure projects with huge current valuations, maintaining sky-high FDVs depends more on the imagination of their narratives. However, the current situation where the exaggerated valuations of infrastructure projects are inconsistent with their actual operating conditions will not only cause the market to fall into the bubble of the "unending prosperity" of the previous bull market, but also the mindset of everyone wanting to earn rent while lying down has brought hidden dangers of the tragedy of the commons to Crypto.