[DSRV ResearchPedia X GAIROS] CLOBs on Blobs: The Evolution of DEX Using Celestia

[DSRV ResearchPedia X GAIROS] CLOBs on Blobs: The Evolution of DEX Using Celestia

[DSRV ResearchPedia X GAIROS] CLOBs on Blobs: The Evolution of DEX Using Celestia

Aug 30, 2025

Aug 30, 2025

💊 Key Takeaways

  • A Celestia-based CLOB DEX offers lower slippage, deeper liquidity, and diverse order strategies compared to traditional AMMs, catering to institutional and professional trader demand.

  • Celestia’s modular DA technologies—DAS, NMT, Blobstream—simultaneously solve issues of speed, scalability, and security, enabling hybrid models like Hibachi that combine CEX-level performance with DEX transparency.

  • As such, Celestia-powered CLOBs elevate professionalism, scalability, and cross-chain interoperability in DeFi, presenting a new trading paradigm.

1. What is a CLOB?

1.1 Basic Concept and Components

A Central Limit Order Book (CLOB) is a trading system where buy and sell orders are recorded in a centralized order book and matched by a matching engine. Participants can submit limit or market orders, and the book sorts bids from highest to lowest and asks from lowest to highest. When bid and ask prices meet, trades execute under price-time priority.

Key components include:

  • Order Book: Records all resting limit orders in real time, providing market depth visibility.

  • Matching Engine: Connects buy and sell orders based on price-time priority.

  • Settlement System: Transfers and clears assets once trades execute (in DeFi, automated via smart contracts).

While traditionally used in CEXs, decentralized CLOB implementations are increasingly being adopted in DeFi.

1.2 AMM vs. CLOB

AMMs (Automated Market Makers) dominate DeFi trading, where LPs deposit assets into pools and pricing follows pre-set formulas. While AMMs allow continuous trading, they suffer from high slippage in large trades, capital inefficiency, divergence from external markets, and impermanent loss for LPs.

CLOBs, by contrast, allow precise pricing, deeper liquidity at specific price levels, and stability for large orders, with advanced order types enabling sophisticated strategies. Institutions and professional traders prefer CLOBs for these reasons, as they reduce fragmentation and replicate familiar trading environments.


2. Why is Celestia Needed to Implement CLOBs?

2.1 The Resurgence of CLOBs and Their Limitations

Central Limit Order Books (CLOBs) are regaining attention due to the maturation of the decentralized finance (DeFi) ecosystem and the entry of institutional investors. Spot trading volumes on decentralized exchanges (DEXs) now exceed 20% of centralized exchange (CEX) volumes, surpassing $450 billion per month. Against this backdrop, the CLOB mechanism, long proven in traditional finance (TradFi), is expanding into the DeFi space. The shifting requirements of institutional investors are the key driver. While Automated Market Makers (AMMs) are suitable for simple swaps, they fall short for institutions that demand more complex and precise trading. Institutions require limited slippage, fine-tuned price control, and advanced order types (such as stop-loss or conditional orders)—capabilities that only CLOBs can provide.

CLOBs offer several core advantages over AMMs:

  • Price discovery efficiency: In a CLOB, prices are determined directly by participants’ buy and sell intentions, whereas AMMs rely on pre-set mathematical formulas. This makes CLOBs more accurate in reflecting true market prices.

  • Capital efficiency: In AMMs, liquidity providers often act as “sitting ducks,” enabling arbitrageurs to extract profits. In CLOBs, market makers actively manage risk and provide deeper, higher-quality liquidity.

  • Variety of trading tools: CLOBs support diverse order types—limit orders, market orders, stop orders, iceberg orders—allowing for sophisticated trading strategies.

However, there are serious technical constraints to implementing CLOBs on-chain.

  • Processing speed: On-chain CLOBs require tens of thousands of transactions per second to operate profitably for market makers, but existing L1 and L2 blockchains cannot meet this demand.

  • Gas fees: The complex operations of a CLOB can cause gas costs to spike under network congestion, pricing out smaller traders.

  • Front-running and MEV: On-chain transparency exposes CLOBs to front-running attacks, where validators can reorder transactions to capture spreads.

  • Scalability limits: Monolithic blockchains handle execution, consensus, and data availability on a single layer, creating the scalability trilemma. For CLOBs, which require both real-time order matching and guaranteed data availability, this architecture falls short.

Because of these limitations, the successful implementation of on-chain CLOBs requires a modular blockchain architecture and a dedicated data availability layer—which is exactly what Celestia provides.

Hyperliquid can process more than 100,000 orders per second, giving it unmatched throughput—no fully on-chain CLOB surpasses it in raw performance. The network focuses on perpetual futures and spot orders, delivering performance once thought impossible on traditional blockchains. However, the project faces significant centralization concerns. Development is conducted as a “black box,” validator power is concentrated among a handful of operators to support high-performance nodes, and the foundation exerts outsized influence over staking and other areas.

In March 2025, a whale investor excessively pumped a token he had created called Jelly. To prevent catastrophic losses in the liquidity pool, Hyperliquid exercised centralized control by rolling back the token’s price. Following a unanimous validator vote, the price oracle was readjusted, resetting Jelly’s value from $0.50 to $0.0095—the price point at which the attacker had initiated short positions. Although the attacker had deposited $7.17 million into his account, his funds were frozen mid-attack, allowing him to withdraw only $6.26 million. He ended up losing roughly $1 million, while the Hyperliquid Liquidity Pool (HLP) paradoxically earned a net profit of $700,000 over 24 hours.

Blockchain security firm Halborn noted that this intervention, while preventing losses, paradoxically damaged Hyperliquid’s credibility. They emphasized the need for comprehensive threat modeling and bug bounty programs during system design and implementation to avoid similar issues.

This illustrates a broader critique of existing on-chain CLOBs: while they deliver fast trading speeds, they often compromise either decentralization or security.

By contrast, Celestia’s core philosophy is the separation of roles, focusing exclusively on ensuring data availability (DA) at the base layer of the blockchain. Functionally, this is somewhat analogous to cloud services that outsource IT resources, but unlike traditional cloud models, Celestia avoids the centralization pitfalls that plague them.

2.2 Celestia’s Technical Differentiators and Their Connection to CLOB Requirements

Celestia’s unique technological advantages can be divided into three categories:

  1. Data Availability Sampling (DAS)

  2. Namespaced Merkle Tree (NMT)

  3. Blobstream

To understand DAS, one must first understand the concept of data availability and how Celestia implements it. This concept is detailed in the 2019 paper LazyLedger (later renamed Celestia) by Mustafa Al-Bassam, Celestia’s co-founder.

| Data availability answers the question: “Has this data actually been published?”

When a node receives a new block added to the chain, it attempts to download all of the block’s transaction data in order to verify data availability. If the node can successfully download the full dataset, the block is considered to have passed data availability verification—meaning the block’s data has indeed been published to the network.

Data availability ensures that anyone can review and validate the transaction ledger, making it a critical element of blockchain security. However, this concept also creates challenges for blockchain scalability. As block sizes grow, it becomes practically impossible for regular users to download all the data, making full chain verification infeasible.

To address this issue, Celestia employs erasure coding, specifically 2D Reed–Solomon (RS) encoding, to implement data availability in a scalable way.


To guarantee data availability, Celestia’s individual nodes sample from a 2k × 2k data matrix. First, the dataset is extended with parity rows and columns (2k each), producing a total of 4k Merkle roots. From this expanded grid, each node randomly selects some coordinate pairs from a “unique coordinate set” for sampling.

With 2D Reed–Solomon encoding, verifying a block header requires generating these 4k Merkle roots. Lightweight nodes only need to sample a few dozen data shares. They then query full nodes for the data corresponding to those sampled coordinates. If valid responses are returned for all queries, the block’s data availability can be guaranteed with very high probability.

In principle, DAS does not have to follow this method. Using a standard approach, one could compute a single Merkle root for the original k data chunks combined with k parity chunks, forming a 2k set. However, Celestia avoids this standard method for a reason: it would require assuming that a majority of consensus participants (block producers) are honest. With Celestia, individual nodes must be able to independently reject invalid blocks. Under the standard approach, this would impose significant overhead, since detecting inconsistency would require downloading the entire original dataset. By contrast, 2D Reed–Solomon encoding allows verification by checking only a single row or column of the extended matrix, making it far more efficient.

A Namespaced Merkle Tree (NMT) is a type of ordered Merkle tree that uses an enhanced hashing function, where each node is sorted according to the namespace range of the messages included in all of its child nodes.

The parent node’s hash is defined as follows:

Hash(child1_minID || child1_maxID || child1_hash || 
     child2_minID || child2_maxID || child2_hash)

Unlike conventional Merkle Trees or Segment Trees, this approach incorporates not just the child node hashes but also the minimum and maximum Namespace IDs of the children. The hash function itself can still be a standard one (e.g., SHA-256), but the inputs include these namespace boundaries.

This design is somewhat similar to how Ethereum externally owned account (EOA) addresses are derived—where an elliptic curve public key is hashed using keccak-256 and the last 20 bytes are taken. Comparable implementations are also found in trie data structures.

Blobstream is the first Ethereum data availability (DA) solution that scales securely with the number of users. In Celestia’s case, the implementation used is SP1 Blobstream, developed by Succinct. Written in Rust, it offers high performance as well as compatibility with EVM-based chains. This means Ethereum developers can leverage Celestia’s optimized Data Availability Sampling (DAS) layer to build high-throughput Layer 2 contracts. Entire ecosystems can also deploy a Blobstream light client on-chain, allowing L2 and L3 contracts to access Celestia’s data availability layer.

Furthermore, Celestia’s Blobstream architecture is designed using RISC Zero and is being released as open source, which will provide significant benefits to the blockchain ecosystem. Not only will dApp performance improve across chains, but rollups and applications will also be able to verify Celestia’s data availability through zero-knowledge (ZK) proofs.

2.3 Improvements in Cost and Speed

When executing Blockvalid(Hi), the bandwidth cost has a time complexity of O(N​+log(N​))

This is because nodes only need to download a 2√N row set from the 2D erasure-coded data, along with the Merkle root, a fixed number of share samples, and the corresponding Merkle proofs.

2.4 Practical Implementation: Integration with Hibachi

Hibachi is a privacy-focused exchange that prioritizes both speed and accuracy, built on a user-first philosophy. Guided by the vision of “following intuition rather than ideology,” its mission is to deliver anonymity and a seamless trading experience to individual users. As discussed earlier in Section 2.1, while CLOB-based DEXs have clear advantages, they have historically been implemented mainly in CEXs, as achieving the necessary transaction speed without centralization has been difficult.

Hibachi’s goal is to provide users with strong security in a DEX CLOB setting—without violating the principles of decentralization. On-chain CLOBs risk exposing sensitive information, such as user balances and positions, creating vulnerabilities. Malicious actors could analyze this data to inflict severe economic losses on individual traders. For example, well-known trader James Wynn reportedly lost $100 million in a single week during leveraged Bitcoin trades on one platform and publicly warned others of the dangers.

As Wynn’s case illustrates, MEV (Maximal Extractable Value) vulnerabilities exist in DEX CLOBs. A common attack is the sandwich attack, where attackers observe large pending transactions in the public mempool, pay higher gas fees to front-run the order, then sell back afterward for profit. Hibachi addresses this trust issue by adopting SP1 zkVM technology, which encodes a First-Come, First-Served (FCFS) policy directly into the Rust-based code. The zkVM then cryptographically proves that the policy was executed exactly as written, preventing MEV exploits.

Other MEV defense approaches exist, such as threshold encryption, which prevents sequencers from seeing transaction contents altogether. This eliminates MEV opportunities since sequencers cannot determine the most profitable ordering. However, Hibachi does not adopt this method. Instead, it allows sequencers to view transactions but enforces the FCFS policy through zkVM proofs. If sequencers attempt to reorder transactions, no valid ZK proof can be generated, and the block cannot be recorded on L1. Rather than a “zero trust” approach, Hibachi chooses a cryptographically auditable sequencer model, consistent with its philosophy of building provable systems.

Although security was the main driver for adopting zkVM, Hibachi also aims to achieve CEX-level transaction speed.

To achieve this, Hibachi presents a hybrid approach, combining Succinct’s advanced ZK technology stack with Celestia’s modular architecture. Succinct’s high-performance zkVM efficiently processes complex logic like order matching, asset management, and risk management on RISC Zero’s zkVM, while Celestia’s data availability layer decentralizes transaction data and ensures verifiability. Together, this successfully merges the speed of CEXs with the transparency of DEXs. Hibachi stands as a strong example of Celestia’s modular blockchain philosophy in practice.

Notably, Succinct’s SP1 zkVM, adopted by Hibachi, also has a modular structure. Built on Plonky3, it preserves the strengths of Plonky2 but completely redesigns the architecture to maximize modularity, performance, and flexibility. Unlike Plonky2, which simply combined PLONK and FRI, Plonky3 modularizes them for flexible composition and supports efficient recursive proofs.

It further incorporates the Sumcheck protocol to maximize arithmetic performance and enables flexible optimization through user-defined custom gates. These features make it well-aligned with the vision of modular blockchains.

3. Conclusion & Outlook: The Significance and Future of Celestia-Based CLOBs

Celestia-based CLOBs overcome the limitations of monolithic blockchains and represent a true turning point in realizing on-chain central order book structures. While AMM-based DEXs have focused on accessibility and automation, CLOBs are optimized to meet the sophisticated trading demands of liquidity concentration, low slippage, and diverse order strategies. This structure is drawing greater attention as institutional and professional traders increasingly expand into on-chain markets. Celestia stands at the core of this structural shift. As a modular blockchain dedicated to data availability (DAS), it solves the challenge of processing large-scale order data that Layer 1 networks struggled with, while achieving scalability and cost efficiency through integration with the rollup ecosystem. In particular, Blobstream and erasure coding-based designs provide both verifiability and reliability for on-chain transactions, enabling a new trading experience that combines DeFi’s transparency with CEX-level speed. A real-world example is Hibachi, which combined Celestia’s data layer with Succinct’s ZK engine to implement a hybrid DEX model that balances privacy protection with high-performance matching. This stands not only as a technical possibility but also as a representative case proving the interoperability, scalability, and practical applicability of modular blockchains.

Outlook

The on-chain entry of institutional investors is becoming increasingly visible. Demand that was once heavily dependent on centralized exchanges (CEXs) is now finding a foundation to move toward Celestia-based CLOBs. This can be interpreted not simply as a technological change, but as a signal accelerating the structural evolution of the DeFi market. At the center of this shift is the modular blockchain. Celestia’s separation of execution, consensus, and data availability introduces a new paradigm for blockchain architecture and generates strong synergies with both the Cosmos and Ethereum rollup ecosystems. This allows developers to flexibly build applications optimized for specific purposes, contributing to greater diversity and sophistication across the ecosystem.

Significant progress is also being made in the balance of privacy and scalability. By integrating zero-knowledge (ZK) technology, users gain protection of balances and positions while preventing exposure of trading strategies, all while maintaining transparency of on-chain settlement. This represents an important evolution beyond the limitations of traditional DEXs.

Additionally, the spread of Blobstream, one of Celestia’s core modules, is also noteworthy. By enabling interoperability with Ethereum and various other chains, Celestia is likely to become the new standard for data availability (DA) in the Layer 2 ecosystem. Ultimately, Celestia-based CLOBs are more than just a technical experiment; they serve as a concrete example of how decentralized exchanges can simultaneously achieve professionalism and scalability. These attempts to merge the speed of centralized exchanges with the transparency of decentralized exchanges are poised to become the practical benchmark of the modular blockchain era, with Celestia positioned at its core.


💊 Key Takeaways

  • A Celestia-based CLOB DEX offers lower slippage, deeper liquidity, and diverse order strategies compared to traditional AMMs, catering to institutional and professional trader demand.

  • Celestia’s modular DA technologies—DAS, NMT, Blobstream—simultaneously solve issues of speed, scalability, and security, enabling hybrid models like Hibachi that combine CEX-level performance with DEX transparency.

  • As such, Celestia-powered CLOBs elevate professionalism, scalability, and cross-chain interoperability in DeFi, presenting a new trading paradigm.

1. What is a CLOB?

1.1 Basic Concept and Components

A Central Limit Order Book (CLOB) is a trading system where buy and sell orders are recorded in a centralized order book and matched by a matching engine. Participants can submit limit or market orders, and the book sorts bids from highest to lowest and asks from lowest to highest. When bid and ask prices meet, trades execute under price-time priority.

Key components include:

  • Order Book: Records all resting limit orders in real time, providing market depth visibility.

  • Matching Engine: Connects buy and sell orders based on price-time priority.

  • Settlement System: Transfers and clears assets once trades execute (in DeFi, automated via smart contracts).

While traditionally used in CEXs, decentralized CLOB implementations are increasingly being adopted in DeFi.

1.2 AMM vs. CLOB

AMMs (Automated Market Makers) dominate DeFi trading, where LPs deposit assets into pools and pricing follows pre-set formulas. While AMMs allow continuous trading, they suffer from high slippage in large trades, capital inefficiency, divergence from external markets, and impermanent loss for LPs.

CLOBs, by contrast, allow precise pricing, deeper liquidity at specific price levels, and stability for large orders, with advanced order types enabling sophisticated strategies. Institutions and professional traders prefer CLOBs for these reasons, as they reduce fragmentation and replicate familiar trading environments.


2. Why is Celestia Needed to Implement CLOBs?

2.1 The Resurgence of CLOBs and Their Limitations

Central Limit Order Books (CLOBs) are regaining attention due to the maturation of the decentralized finance (DeFi) ecosystem and the entry of institutional investors. Spot trading volumes on decentralized exchanges (DEXs) now exceed 20% of centralized exchange (CEX) volumes, surpassing $450 billion per month. Against this backdrop, the CLOB mechanism, long proven in traditional finance (TradFi), is expanding into the DeFi space. The shifting requirements of institutional investors are the key driver. While Automated Market Makers (AMMs) are suitable for simple swaps, they fall short for institutions that demand more complex and precise trading. Institutions require limited slippage, fine-tuned price control, and advanced order types (such as stop-loss or conditional orders)—capabilities that only CLOBs can provide.

CLOBs offer several core advantages over AMMs:

  • Price discovery efficiency: In a CLOB, prices are determined directly by participants’ buy and sell intentions, whereas AMMs rely on pre-set mathematical formulas. This makes CLOBs more accurate in reflecting true market prices.

  • Capital efficiency: In AMMs, liquidity providers often act as “sitting ducks,” enabling arbitrageurs to extract profits. In CLOBs, market makers actively manage risk and provide deeper, higher-quality liquidity.

  • Variety of trading tools: CLOBs support diverse order types—limit orders, market orders, stop orders, iceberg orders—allowing for sophisticated trading strategies.

However, there are serious technical constraints to implementing CLOBs on-chain.

  • Processing speed: On-chain CLOBs require tens of thousands of transactions per second to operate profitably for market makers, but existing L1 and L2 blockchains cannot meet this demand.

  • Gas fees: The complex operations of a CLOB can cause gas costs to spike under network congestion, pricing out smaller traders.

  • Front-running and MEV: On-chain transparency exposes CLOBs to front-running attacks, where validators can reorder transactions to capture spreads.

  • Scalability limits: Monolithic blockchains handle execution, consensus, and data availability on a single layer, creating the scalability trilemma. For CLOBs, which require both real-time order matching and guaranteed data availability, this architecture falls short.

Because of these limitations, the successful implementation of on-chain CLOBs requires a modular blockchain architecture and a dedicated data availability layer—which is exactly what Celestia provides.

Hyperliquid can process more than 100,000 orders per second, giving it unmatched throughput—no fully on-chain CLOB surpasses it in raw performance. The network focuses on perpetual futures and spot orders, delivering performance once thought impossible on traditional blockchains. However, the project faces significant centralization concerns. Development is conducted as a “black box,” validator power is concentrated among a handful of operators to support high-performance nodes, and the foundation exerts outsized influence over staking and other areas.

In March 2025, a whale investor excessively pumped a token he had created called Jelly. To prevent catastrophic losses in the liquidity pool, Hyperliquid exercised centralized control by rolling back the token’s price. Following a unanimous validator vote, the price oracle was readjusted, resetting Jelly’s value from $0.50 to $0.0095—the price point at which the attacker had initiated short positions. Although the attacker had deposited $7.17 million into his account, his funds were frozen mid-attack, allowing him to withdraw only $6.26 million. He ended up losing roughly $1 million, while the Hyperliquid Liquidity Pool (HLP) paradoxically earned a net profit of $700,000 over 24 hours.

Blockchain security firm Halborn noted that this intervention, while preventing losses, paradoxically damaged Hyperliquid’s credibility. They emphasized the need for comprehensive threat modeling and bug bounty programs during system design and implementation to avoid similar issues.

This illustrates a broader critique of existing on-chain CLOBs: while they deliver fast trading speeds, they often compromise either decentralization or security.

By contrast, Celestia’s core philosophy is the separation of roles, focusing exclusively on ensuring data availability (DA) at the base layer of the blockchain. Functionally, this is somewhat analogous to cloud services that outsource IT resources, but unlike traditional cloud models, Celestia avoids the centralization pitfalls that plague them.

2.2 Celestia’s Technical Differentiators and Their Connection to CLOB Requirements

Celestia’s unique technological advantages can be divided into three categories:

  1. Data Availability Sampling (DAS)

  2. Namespaced Merkle Tree (NMT)

  3. Blobstream

To understand DAS, one must first understand the concept of data availability and how Celestia implements it. This concept is detailed in the 2019 paper LazyLedger (later renamed Celestia) by Mustafa Al-Bassam, Celestia’s co-founder.

| Data availability answers the question: “Has this data actually been published?”

When a node receives a new block added to the chain, it attempts to download all of the block’s transaction data in order to verify data availability. If the node can successfully download the full dataset, the block is considered to have passed data availability verification—meaning the block’s data has indeed been published to the network.

Data availability ensures that anyone can review and validate the transaction ledger, making it a critical element of blockchain security. However, this concept also creates challenges for blockchain scalability. As block sizes grow, it becomes practically impossible for regular users to download all the data, making full chain verification infeasible.

To address this issue, Celestia employs erasure coding, specifically 2D Reed–Solomon (RS) encoding, to implement data availability in a scalable way.


To guarantee data availability, Celestia’s individual nodes sample from a 2k × 2k data matrix. First, the dataset is extended with parity rows and columns (2k each), producing a total of 4k Merkle roots. From this expanded grid, each node randomly selects some coordinate pairs from a “unique coordinate set” for sampling.

With 2D Reed–Solomon encoding, verifying a block header requires generating these 4k Merkle roots. Lightweight nodes only need to sample a few dozen data shares. They then query full nodes for the data corresponding to those sampled coordinates. If valid responses are returned for all queries, the block’s data availability can be guaranteed with very high probability.

In principle, DAS does not have to follow this method. Using a standard approach, one could compute a single Merkle root for the original k data chunks combined with k parity chunks, forming a 2k set. However, Celestia avoids this standard method for a reason: it would require assuming that a majority of consensus participants (block producers) are honest. With Celestia, individual nodes must be able to independently reject invalid blocks. Under the standard approach, this would impose significant overhead, since detecting inconsistency would require downloading the entire original dataset. By contrast, 2D Reed–Solomon encoding allows verification by checking only a single row or column of the extended matrix, making it far more efficient.

A Namespaced Merkle Tree (NMT) is a type of ordered Merkle tree that uses an enhanced hashing function, where each node is sorted according to the namespace range of the messages included in all of its child nodes.

The parent node’s hash is defined as follows:

Hash(child1_minID || child1_maxID || child1_hash || 
     child2_minID || child2_maxID || child2_hash)

Unlike conventional Merkle Trees or Segment Trees, this approach incorporates not just the child node hashes but also the minimum and maximum Namespace IDs of the children. The hash function itself can still be a standard one (e.g., SHA-256), but the inputs include these namespace boundaries.

This design is somewhat similar to how Ethereum externally owned account (EOA) addresses are derived—where an elliptic curve public key is hashed using keccak-256 and the last 20 bytes are taken. Comparable implementations are also found in trie data structures.

Blobstream is the first Ethereum data availability (DA) solution that scales securely with the number of users. In Celestia’s case, the implementation used is SP1 Blobstream, developed by Succinct. Written in Rust, it offers high performance as well as compatibility with EVM-based chains. This means Ethereum developers can leverage Celestia’s optimized Data Availability Sampling (DAS) layer to build high-throughput Layer 2 contracts. Entire ecosystems can also deploy a Blobstream light client on-chain, allowing L2 and L3 contracts to access Celestia’s data availability layer.

Furthermore, Celestia’s Blobstream architecture is designed using RISC Zero and is being released as open source, which will provide significant benefits to the blockchain ecosystem. Not only will dApp performance improve across chains, but rollups and applications will also be able to verify Celestia’s data availability through zero-knowledge (ZK) proofs.

2.3 Improvements in Cost and Speed

When executing Blockvalid(Hi), the bandwidth cost has a time complexity of O(N​+log(N​))

This is because nodes only need to download a 2√N row set from the 2D erasure-coded data, along with the Merkle root, a fixed number of share samples, and the corresponding Merkle proofs.

2.4 Practical Implementation: Integration with Hibachi

Hibachi is a privacy-focused exchange that prioritizes both speed and accuracy, built on a user-first philosophy. Guided by the vision of “following intuition rather than ideology,” its mission is to deliver anonymity and a seamless trading experience to individual users. As discussed earlier in Section 2.1, while CLOB-based DEXs have clear advantages, they have historically been implemented mainly in CEXs, as achieving the necessary transaction speed without centralization has been difficult.

Hibachi’s goal is to provide users with strong security in a DEX CLOB setting—without violating the principles of decentralization. On-chain CLOBs risk exposing sensitive information, such as user balances and positions, creating vulnerabilities. Malicious actors could analyze this data to inflict severe economic losses on individual traders. For example, well-known trader James Wynn reportedly lost $100 million in a single week during leveraged Bitcoin trades on one platform and publicly warned others of the dangers.

As Wynn’s case illustrates, MEV (Maximal Extractable Value) vulnerabilities exist in DEX CLOBs. A common attack is the sandwich attack, where attackers observe large pending transactions in the public mempool, pay higher gas fees to front-run the order, then sell back afterward for profit. Hibachi addresses this trust issue by adopting SP1 zkVM technology, which encodes a First-Come, First-Served (FCFS) policy directly into the Rust-based code. The zkVM then cryptographically proves that the policy was executed exactly as written, preventing MEV exploits.

Other MEV defense approaches exist, such as threshold encryption, which prevents sequencers from seeing transaction contents altogether. This eliminates MEV opportunities since sequencers cannot determine the most profitable ordering. However, Hibachi does not adopt this method. Instead, it allows sequencers to view transactions but enforces the FCFS policy through zkVM proofs. If sequencers attempt to reorder transactions, no valid ZK proof can be generated, and the block cannot be recorded on L1. Rather than a “zero trust” approach, Hibachi chooses a cryptographically auditable sequencer model, consistent with its philosophy of building provable systems.

Although security was the main driver for adopting zkVM, Hibachi also aims to achieve CEX-level transaction speed.

To achieve this, Hibachi presents a hybrid approach, combining Succinct’s advanced ZK technology stack with Celestia’s modular architecture. Succinct’s high-performance zkVM efficiently processes complex logic like order matching, asset management, and risk management on RISC Zero’s zkVM, while Celestia’s data availability layer decentralizes transaction data and ensures verifiability. Together, this successfully merges the speed of CEXs with the transparency of DEXs. Hibachi stands as a strong example of Celestia’s modular blockchain philosophy in practice.

Notably, Succinct’s SP1 zkVM, adopted by Hibachi, also has a modular structure. Built on Plonky3, it preserves the strengths of Plonky2 but completely redesigns the architecture to maximize modularity, performance, and flexibility. Unlike Plonky2, which simply combined PLONK and FRI, Plonky3 modularizes them for flexible composition and supports efficient recursive proofs.

It further incorporates the Sumcheck protocol to maximize arithmetic performance and enables flexible optimization through user-defined custom gates. These features make it well-aligned with the vision of modular blockchains.

3. Conclusion & Outlook: The Significance and Future of Celestia-Based CLOBs

Celestia-based CLOBs overcome the limitations of monolithic blockchains and represent a true turning point in realizing on-chain central order book structures. While AMM-based DEXs have focused on accessibility and automation, CLOBs are optimized to meet the sophisticated trading demands of liquidity concentration, low slippage, and diverse order strategies. This structure is drawing greater attention as institutional and professional traders increasingly expand into on-chain markets. Celestia stands at the core of this structural shift. As a modular blockchain dedicated to data availability (DAS), it solves the challenge of processing large-scale order data that Layer 1 networks struggled with, while achieving scalability and cost efficiency through integration with the rollup ecosystem. In particular, Blobstream and erasure coding-based designs provide both verifiability and reliability for on-chain transactions, enabling a new trading experience that combines DeFi’s transparency with CEX-level speed. A real-world example is Hibachi, which combined Celestia’s data layer with Succinct’s ZK engine to implement a hybrid DEX model that balances privacy protection with high-performance matching. This stands not only as a technical possibility but also as a representative case proving the interoperability, scalability, and practical applicability of modular blockchains.

Outlook

The on-chain entry of institutional investors is becoming increasingly visible. Demand that was once heavily dependent on centralized exchanges (CEXs) is now finding a foundation to move toward Celestia-based CLOBs. This can be interpreted not simply as a technological change, but as a signal accelerating the structural evolution of the DeFi market. At the center of this shift is the modular blockchain. Celestia’s separation of execution, consensus, and data availability introduces a new paradigm for blockchain architecture and generates strong synergies with both the Cosmos and Ethereum rollup ecosystems. This allows developers to flexibly build applications optimized for specific purposes, contributing to greater diversity and sophistication across the ecosystem.

Significant progress is also being made in the balance of privacy and scalability. By integrating zero-knowledge (ZK) technology, users gain protection of balances and positions while preventing exposure of trading strategies, all while maintaining transparency of on-chain settlement. This represents an important evolution beyond the limitations of traditional DEXs.

Additionally, the spread of Blobstream, one of Celestia’s core modules, is also noteworthy. By enabling interoperability with Ethereum and various other chains, Celestia is likely to become the new standard for data availability (DA) in the Layer 2 ecosystem. Ultimately, Celestia-based CLOBs are more than just a technical experiment; they serve as a concrete example of how decentralized exchanges can simultaneously achieve professionalism and scalability. These attempts to merge the speed of centralized exchanges with the transparency of decentralized exchanges are poised to become the practical benchmark of the modular blockchain era, with Celestia positioned at its core.


BD Manager

Youngbin Park

Written by

Youngbin Park

Aug 30, 2025

© 2025. DSRV labs. All rights reserved

© 2025. DSRV labs. All rights reserved

DSRV, 73, Teheran-ro 19-gil,
Gangnam-gu, Seoul, Republic of Korea

© 2025. DSRV labs. All rights reserved