Taiko CTO Says Most Rollups Aren’t Really Decentralized—Here’s Why

Brecht Devos, the Co-Founder and CTO of Taiko, sits at the forefront of Ethereum’s scaling revolution with a rollup designed to uphold its core values of decentralization and permissionlessness. While many rollups prioritize speed and centralization, Taiko aims to mirror Ethereum as closely as possible, ensuring real-time censorship resistance and composability. In this exclusive interview, Brecht shares the inspiration behind Taiko’s innovative approach, the groundbreaking technologies driving its growth, and how it envisions reshaping Ethereum’s future amidst the proliferation of Layer 2 solutions like zkSync and Arbitrum. From the implementation of multi-proof architecture to the transformative potential of transaction preconfirmation, this conversation delves deep into the mechanics, challenges, and vision behind Taiko’s mission to make Ethereum scalable without compromising its foundational principles. Brecht, as the Co-Founder and CTO of Taiko, what inspired you to build a rollup that prioritizes Ethereum’s core values of decentralization and permissionlessness? The inspiration for Taiko was to build a rollup that resembled Ethereum as closely as possible. We had noticed that while there were many rollups out there, not many prioritized decentralization but rather speed, resulting in centralized sequencers that provide a permissioned point of failure. Permissionless is really what being based provides. This can guarantee real-time censorship resistance (or at least as good as Ethereum's) and strong liveness guarantees. Even more, it’s credibly neutral which is important for composability. For most practical purposes, being based and thereby permissionless is just to achieve the goal of being maximally composable with L1 and other L2s to create a scalable Ethereum that feels like a single chain, with cheap fees. With Ethereum’s rollup-centric roadmap, rollups are part of Ethereum’s future and so need to align with its core values. With Taiko being the first "based rollup," how do you envision its role in shaping the future of Ethereum's ecosystem, particularly as Layer 2 solutions proliferate? As the pioneer, Taiko sets the standard for what is possible. Now that other projects can see it is viable to run a rollup with native Ethereum sequencers, it’s likely more will follow and launch their own. So, we’ll likely see more based rollups this year and moving forward, with many of those not simply copying our formula but making tweaks to improve and strengthen their capabilities, like we’re consistently doing at Taiko. We hope that in the future, these L2s will work together to give the best user experience possible rather than competing. With more L2s looking to be based, our goals of scaling Ethereum in the best way possible are almost automatically aligned, and so the future here seems bright. Taiko's multi-proof architecture, powered by Succinct and RISC Zero, is a groundbreaking feature. Can you explain how this architecture works and why it is a game-changer for Layer 2 scaling? A multi-proof system increases security, as it aims to address any possible bugs or concerns that could be overlooked in a single-proof system. With multiple provers, there is more oversight and it’s unlikely that if one prover has a bug or an issue that the other prover has the exact same problem. Provers can work very differently not just on what is being proven (similar to the multi-client approach of Ethereum) but also how it is being proven (TEE vs ZK proofs). Combining these different approaches can make a very secure permissionless system. For the long-term, based rollups especially benefit from multi-proof systems because blocks are not proposed by a trusted centralized sequencer that would not want to exploit vulnerabilities normally. As more rollups consider based architecture, these security measures need to be considered as there is little point in having decentralization without effective security. Processing 5 million daily transactions is no small feat. How does Taiko ensure security and decentralization while maintaining such high throughput? Decentralization is ensured by having a higher but very reasonable throughput limitation for the Taiko rollup. A problem for decentralization would be if you need expensive hardware to run a Taiko node yourself. The throughput on Taiko is limited the same way it is on Ethereum itself through EIP-1559. Once the throughput gets too high, the fee goes up until the throughput returns to normal. Decentralization/Permissionless is also guaranteed through being a based rollup where anybody can propose/prove a block, with each block needing to be proven by TEE and, soon, at least one ZK proof. We are working on adding additional provers to increase the security further. Transaction preconfirmation technology is set to revolutionize the user experience. How does this bypass Ethereum’s block confirmation times, and what are the implications for mass adoption? Preconfirmations will be a huge user experience benefit on based rollups and Ethereum L1. It will allow users to have the same expected user experience as they do for any other application on the internet or web2 apps. This means no more uncertainty around when your transaction will get executed and what the outcome will be. Currently, preconfirmations are achieved through an out-of-protocol system that allows users to directly (or indirectly) interact with the L1 validator who proposes the next block which will already guarantee the inclusion of a transaction in its block before it is eventually made a part of the Ethereum blockchain. What challenges did you encounter in implementing transaction preconfirmation, and how did the team overcome them? Preconfirmations, while quite simple in design, were a huge undertaking. It took a lot of the Ethereum ecosystem coming together to solve this problem over roughly a year to get where we are today, and there’s still work to be done. But the sheer amount of energy and drive from a whole range of projects and people made it possible to now actually start using them. We would not have been able to do this without this shared effort, and Justin Drake leading it. The design is relatively simple but in practice it is surprisingly complex for several reasons. Building them required us having to go through the many actors in the Ethereum block-building pipeline, starting from the wallet all the way to the L1 validator. Collaboration was thus key here, and Taiko being the only based rollup launched and available for use contributed significantly. The Ontake Hardfork brought significant upgrades like enhanced EIP-1559 mechanisms and batch transaction support. Can you elaborate on how these changes optimize Taiko’s efficiency and user affordability? The biggest changes for efficiency are the support for preconfirmation and proof aggregation. When tuned correctly, preconfirmations will allow more transactions to be batched in a block, decreasing the onchain gas costs. Proof aggregation allows multiple blocks to be proven in a single proof that is verified on-chain, compared to before, when a separate proof needed to be verified on-chain for each block individually. Allowing the EIP-1559 parameters to be more easily tuned is important for controlling the throughput and thus the base fee. Generally, a higher throughput lowers the base fee and the user costs. Still, a higher throughput can also be bad for decentralization as it can increase hardware requirements for people running a Taiko node. So a good balance needs to be found, while node optimizations are also done so that throughput can be increased without increasing hardware requirements. How does the Ontake Hardfork lay the foundation for Taiko’s upcoming DAO launch and broader ecosystem adoption? Preconfirmations are the biggest update allowing a vastly improved user experience for all the applications on top of Taiko. Now applications that require fast confirmation times can be launched and used on our rollup. The range of optimizations are of course also important from a cost perspective. These allow blocks to be proposed cheaper, which lowers transaction fees for users which is always a good thing. On the DAO side, more and more things will be able to be changed onchain, which will eventually allow the DAO to control them. This furthers our efforts to ensure Taiko is decentralized from the foundations up. Taiko is deeply aligned with Ethereum’s ethos. How do you see Taiko contributing to Ethereum’s long-term vision for scalability and decentralization? At Taiko, we’re not simply looking at what is next, but we have a long-term vision for Ethereum’s development. This is why we have two angles we’re working on. There is Taiko mainnet and then there is Gwyneth. Taiko is a more traditional rollup, with the added benefit of being based. It is addressing the needs of the community right now and we’re always working to ensure it aligns as closely with Ethereum where we are today and over the next few years. Gywneth is our research arm working on the endgame of based native rollups. Our focus here is on Ethereum’s endgame, a focus on modularity and scalability without compromising decentralization. As such, instead of having asynchronous bridging between L1 and L2, we designed for synchronous calls across L1 and L2. This means developers can execute calls from L1 to L2 without waiting, which makes bridging between layers much more straightforward. Essentially, Gwyneth is trying to minimize the amount of friction needed to work across chains, realizing the vision of an Ethereum built on rollups that still feels like a single chain for users. With your extensive experience in zero-knowledge proofs, how do you see ZK technology evolving in Ethereum scaling over the next five years? From a developer perspective, with the rise of general zkVMs we are pretty much already at the end goal of how to write software that can be proven using zk proofs. Now, these zkVMs need to be optimized so that proving gets as cheap and as fast as possible. Cheap is important for obvious reasons, but maybe as important is being able to generate them very fast. In the next year or two we’ll likely have zk proofs that can prove things extremely fast, which opens the door to using them in even more scenarios. The added benefit is that it greatly simplifies how implementations can work, because actions can be verified directly when needed instead of having to work around a proof only being available at some later time. Of course, rollups and privacy applications are already taking advantage of them despite the current limitations. In the future, we’ll also see rollups, like Gwyneth, taking advantage of the very fast proving times to have very fast onchain finality times and eventually enable improvements like synchronous composability with L1 and between rollups. On a longer time frame in Ethereum, it will also enable improvements on the consensus layer for signature aggregation, and improvements on the execution layer in the form of native rollups and the snarkification of the whole block execution which will enable much higher gas limits. What lessons from your time at Loopring and Ethereum’s Privacy and Scaling Explorations team have been most influential in developing Taiko? At Loopring I learned a lot about what it actually means to build a rollup that people want to use. Loopring eventually got all the basic functionality right for payments, trading and NFTs. But, developers were always asking for more and wanted to deploy their own smart contracts. So building a rollup where developers could just deploy their existing smart contracts without changes was very important. While working with the PSE team, I learned a lot more about how complex applications can be built using the now old school way of writing circuits for ZK, while also learning the nitty gritty details of how Ethereum works. The PSE team, together with Scroll at the time, were working on the community ZK-EVM project that had a lot of very smart people who I learnt a lot from. The other important thing I learned is to always try to build for what will be possible, not what is currently possible. The tech moves so quickly that you need to build towards what you think things will look like when you are actually ready to launch. With transaction preconfirmation on the horizon for Q1 2025, what other innovations can users and developers expect from Taiko in the near future? The beginning of 2025 is all in on preconfirmations and eventually our DAO launch. These will further Taiko’s aims to become a more effective and decentralized rollup. From Gwyneth, we are working on building the development infrastructure. We’re excited to get some of the infrastructure up on testnet at the beginning of this year and launch it later in the year. It will be very busy, and we are excited to further discuss what is possible for scaling Ethereum. As Taiko prepares for its DAO launch, what steps are you taking to ensure community-driven governance aligns with the project’s ethos? Ahead of the DAO launch, we will be introducing its architecture to our community to hear feedback and insights following the launch of the DAO on Helkla testnet. The build of our DAO is to ensure that Taiko is decentralized from the foundations all the way through - including how decisions are made. Brecht Devos, the Co-Founder and CTO of Taiko , sits at the forefront of Ethereum’s scaling revolution with a rollup designed to uphold its core values of decentralization and permissionlessness. While many rollups prioritize speed and centralization, Taiko aims to mirror Ethereum as closely as possible, ensuring real-time censorship resistance and composability. Taiko In this exclusive interview, Brecht shares the inspiration behind Taiko’s innovative approach, the groundbreaking technologies driving its growth, and how it envisions reshaping Ethereum’s future amidst the proliferation of Layer 2 solutions like zkSync and Arbitrum. From the implementation of multi-proof architecture to the transformative potential of transaction preconfirmation, this conversation delves deep into the mechanics, challenges, and vision behind Taiko’s mission to make Ethereum scalable without compromising its foundational principles. Brecht, as the Co-Founder and CTO of Taiko, what inspired you to build a rollup that prioritizes Ethereum’s core values of decentralization and permissionlessness? Brecht, as the Co-Founder and CTO of Taiko, what inspired you to build a rollup that prioritizes Ethereum’s core values of decentralization and permissionlessness? Brecht, as the Co-Founder and CTO of Taiko, what inspired you to build a rollup that prioritizes Ethereum’s core values of decentralization and permissionlessness? Brecht, as the Co-Founder and CTO of Taiko, what inspired you to build a rollup that prioritizes Ethereum’s core values of decentralization and permissionlessness? The inspiration for Taiko was to build a rollup that resembled Ethereum as closely as possible. We had noticed that while there were many rollups out there, not many prioritized decentralization but rather speed, resulting in centralized sequencers that provide a permissioned point of failure. Permissionless is really what being based provides. This can guarantee real-time censorship resistance (or at least as good as Ethereum's) and strong liveness guarantees. Even more, it’s credibly neutral which is important for composability. For most practical purposes, being based and thereby permissionless is just to achieve the goal of being maximally composable with L1 and other L2s to create a scalable Ethereum that feels like a single chain, with cheap fees. With Ethereum’s rollup-centric roadmap, rollups are part of Ethereum’s future and so need to align with its core values. With Taiko being the first "based rollup," how do you envision its role in shaping the future of Ethereum's ecosystem, particularly as Layer 2 solutions proliferate? With Taiko being the first "based rollup," how do you envision its role in shaping the future of Ethereum's ecosystem, particularly as Layer 2 solutions proliferate? With Taiko being the first "based rollup," how do you envision its role in shaping the future of Ethereum's ecosystem, particularly as Layer 2 solutions proliferate? With Taiko being the first "based rollup," how do you envision its role in shaping the future of Ethereum's ecosystem, particularly as Layer 2 solutions proliferate? As the pioneer, Taiko sets the standard for what is possible. Now that other projects can see it is viable to run a rollup with native Ethereum sequencers, it’s likely more will follow and launch their own. So, we’ll likely see more based rollups this year and moving forward, with many of those not simply copying our formula but making tweaks to improve and strengthen their capabilities, like we’re consistently doing at Taiko. We hope that in the future, these L2s will work together to give the best user experience possible rather than competing. With more L2s looking to be based, our goals of scaling Ethereum in the best way possible are almost automatically aligned, and so the future here seems bright. Taiko's multi-proof architecture, powered by Succinct and RISC Zero, is a groundbreaking feature. Can you explain how this architecture works and why it is a game-changer for Layer 2 scaling? Taiko's multi-proof architecture, powered by Succinct and RISC Zero, is a groundbreaking feature. Can you explain how this architecture works and why it is a game-changer for Layer 2 scaling? Taiko's multi-proof architecture, powered by Succinct and RISC Zero, is a groundbreaking feature. Can you explain how this architecture works and why it is a game-changer for Layer 2 scaling? Taiko's multi-proof architecture, powered by Succinct and RISC Zero, is a groundbreaking feature. Can you explain how this architecture works and why it is a game-changer for Layer 2 scaling? A multi-proof system increases security, as it aims to address any possible bugs or concerns that could be overlooked in a single-proof system. With multiple provers, there is more oversight and it’s unlikely that if one prover has a bug or an issue that the other prover has the exact same problem. Provers can work very differently not just on what is being proven (similar to the multi-client approach of Ethereum) but also how it is being proven (TEE vs ZK proofs). Combining these different approaches can make a very secure permissionless system. For the long-term, based rollups especially benefit from multi-proof systems because blocks are not proposed by a trusted centralized sequencer that would not want to exploit vulnerabilities normally. As more rollups consider based architecture, these security measures need to be considered as there is little point in having decentralization without effective security. Processing 5 million daily transactions is no small feat. How does Taiko ensure security and decentralization while maintaining such high throughput? Processing 5 million daily transactions is no small feat. How does Taiko ensure security and decentralization while maintaining such high throughput? Processing 5 million daily transactions is no small feat. How does Taiko ensure security and decentralization while maintaining such high throughput? Processing 5 million daily transactions is no small feat. How does Taiko ensure security and decentralization while maintaining such high throughput? Decentralization is ensured by having a higher but very reasonable throughput limitation for the Taiko rollup. A problem for decentralization would be if you need expensive hardware to run a Taiko node yourself. The throughput on Taiko is limited the same way it is on Ethereum itself through EIP-1559. Once the throughput gets too high, the fee goes up until the throughput returns to normal. Decentralization/Permissionless is also guaranteed through being a based rollup where anybody can propose/prove a block, with each block needing to be proven by TEE and, soon, at least one ZK proof. We are working on adding additional provers to increase the security further. Transaction preconfirmation technology is set to revolutionize the user experience. How does this bypass Ethereum’s block confirmation times, and what are the implications for mass adoption? Transaction preconfirmation technology is set to revolutionize the user experience. How does this bypass Ethereum’s block confirmation times, and what are the implications for mass adoption? Transaction preconfirmation technology is set to revolutionize the user experience. How does this bypass Ethereum’s block confirmation times, and what are the implications for mass adoption? Transaction preconfirmation technology is set to revolutionize the user experience. How does this bypass Ethereum’s block confirmation times, and what are the implications for mass adoption? Preconfirmations will be a huge user experience benefit on based rollups and Ethereum L1. It will allow users to have the same expected user experience as they do for any other application on the internet or web2 apps. This means no more uncertainty around when your transaction will get executed and what the outcome will be. Currently, preconfirmations are achieved through an out-of-protocol system that allows users to directly (or indirectly) interact with the L1 validator who proposes the next block which will already guarantee the inclusion of a transaction in its block before it is eventually made a part of the Ethereum blockchain. What challenges did you encounter in implementing transaction preconfirmation, and how did the team overcome them? What challenges did you encounter in implementing transaction preconfirmation, and how did the team overcome them? What challenges did you encounter in implementing transaction preconfirmation, and how did the team overcome them? Preconfirmations, while quite simple in design, were a huge undertaking. It took a lot of the Ethereum ecosystem coming together to solve this problem over roughly a year to get where we are today, and there’s still work to be done. But the sheer amount of energy and drive from a whole range of projects and people made it possible to now actually start using them. We would not have been able to do this without this shared effort, and Justin Drake leading it. The design is relatively simple but in practice it is surprisingly complex for several reasons. Building them required us having to go through the many actors in the Ethereum block-building pipeline, starting from the wallet all the way to the L1 validator. Collaboration was thus key here, and Taiko being the only based rollup launched and available for use contributed significantly. The Ontake Hardfork brought significant upgrades like enhanced EIP-1559 mechanisms and batch transaction support. Can you elaborate on how these changes optimize Taiko’s efficiency and user affordability? The Ontake Hardfork brought significant upgrades like enhanced EIP-1559 mechanisms and batch transaction support. Can you elaborate on how these changes optimize Taiko’s efficiency and user affordability? The Ontake Hardfork brought significant upgrades like enhanced EIP-1559 mechanisms and batch transaction support. Can you elaborate on how these changes optimize Taiko’s efficiency and user affordability? The Ontake Hardfork brought significant upgrades like enhanced EIP-1559 mechanisms and batch transaction support. Can you elaborate on how these changes optimize Taiko’s efficiency and user affordability? The biggest changes for efficiency are the support for preconfirmation and proof aggregation. When tuned correctly, preconfirmations will allow more transactions to be batched in a block, decreasing the onchain gas costs. Proof aggregation allows multiple blocks to be proven in a single proof that is verified on-chain, compared to before, when a separate proof needed to be verified on-chain for each block individually. Allowing the EIP-1559 parameters to be more easily tuned is important for controlling the throughput and thus the base fee. Generally, a higher throughput lowers the base fee and the user costs. Still, a higher throughput can also be bad for decentralization as it can increase hardware requirements for people running a Taiko node. So a good balance needs to be found, while node optimizations are also done so that throughput can be increased without increasing hardware requirements. How does the Ontake Hardfork lay the foundation for Taiko’s upcoming DAO launch and broader ecosystem adoption? How does the Ontake Hardfork lay the foundation for Taiko’s upcoming DAO launch and broader ecosystem adoption? How does the Ontake Hardfork lay the foundation for Taiko’s upcoming DAO launch and broader ecosystem adoption? How does the Ontake Hardfork lay the foundation for Taiko’s upcoming DAO launch and broader ecosystem adoption? Preconfirmations are the biggest update allowing a vastly improved user experience for all the applications on top of Taiko. Now applications that require fast confirmation times can be launched and used on our rollup. The range of optimizations are of course also important from a cost perspective. These allow blocks to be proposed cheaper, which lowers transaction fees for users which is always a good thing. On the DAO side, more and more things will be able to be changed onchain, which will eventually allow the DAO to control them. This furthers our efforts to ensure Taiko is decentralized from the foundations up. Taiko is deeply aligned with Ethereum’s ethos. How do you see Taiko contributing to Ethereum’s long-term vision for scalability and decentralization? Taiko is deeply aligned with Ethereum’s ethos. How do you see Taiko contributing to Ethereum’s long-term vision for scalability and decentralization? Taiko is deeply aligned with Ethereum’s ethos. How do you see Taiko contributing to Ethereum’s long-term vision for scalability and decentralization? At Taiko, we’re not simply looking at what is next, but we have a long-term vision for Ethereum’s development. This is why we have two angles we’re working on. There is Taiko mainnet and then there is Gwyneth. Taiko is a more traditional rollup, with the added benefit of being based. It is addressing the needs of the community right now and we’re always working to ensure it aligns as closely with Ethereum where we are today and over the next few years. Gywneth is our research arm working on the endgame of based native rollups. Our focus here is on Ethereum’s endgame, a focus on modularity and scalability without compromising decentralization. As such, instead of having asynchronous bridging between L1 and L2, we designed for synchronous calls across L1 and L2. This means developers can execute calls from L1 to L2 without waiting, which makes bridging between layers much more straightforward. Essentially, Gwyneth is trying to minimize the amount of friction needed to work across chains, realizing the vision of an Ethereum built on rollups that still feels like a single chain for users. With your extensive experience in zero-knowledge proofs, how do you see ZK technology evolving in Ethereum scaling over the next five years? With your extensive experience in zero-knowledge proofs, how do you see ZK technology evolving in Ethereum scaling over the next five years? With your extensive experience in zero-knowledge proofs, how do you see ZK technology evolving in Ethereum scaling over the next five years? With your extensive experience in zero-knowledge proofs, how do you see ZK technology evolving in Ethereum scaling over the next five years? From a developer perspective, with the rise of general zkVMs we are pretty much already at the end goal of how to write software that can be proven using zk proofs. Now, these zkVMs need to be optimized so that proving gets as cheap and as fast as possible. Cheap is important for obvious reasons, but maybe as important is being able to generate them very fast. In the next year or two we’ll likely have zk proofs that can prove things extremely fast, which opens the door to using them in even more scenarios. The added benefit is that it greatly simplifies how implementations can work, because actions can be verified directly when needed instead of having to work around a proof only being available at some later time. Of course, rollups and privacy applications are already taking advantage of them despite the current limitations. In the future, we’ll also see rollups, like Gwyneth, taking advantage of the very fast proving times to have very fast onchain finality times and eventually enable improvements like synchronous composability with L1 and between rollups. On a longer time frame in Ethereum, it will also enable improvements on the consensus layer for signature aggregation, and improvements on the execution layer in the form of native rollups and the snarkification of the whole block execution which will enable much higher gas limits. What lessons from your time at Loopring and Ethereum’s Privacy and Scaling Explorations team have been most influential in developing Taiko? What lessons from your time at Loopring and Ethereum’s Privacy and Scaling Explorations team have been most influential in developing Taiko? What lessons from your time at Loopring and Ethereum’s Privacy and Scaling Explorations team have been most influential in developing Taiko? What lessons from your time at Loopring and Ethereum’s Privacy and Scaling Explorations team have been most influential in developing Taiko? At Loopring I learned a lot about what it actually means to build a rollup that people want to use. Loopring eventually got all the basic functionality right for payments, trading and NFTs. But, developers were always asking for more and wanted to deploy their own smart contracts. So building a rollup where developers could just deploy their existing smart contracts without changes was very important. While working with the PSE team, I learned a lot more about how complex applications can be built using the now old school way of writing circuits for ZK, while also learning the nitty gritty details of how Ethereum works. The PSE team, together with Scroll at the time, were working on the community ZK-EVM project that had a lot of very smart people who I learnt a lot from. The other important thing I learned is to always try to build for what will be possible, not what is currently possible. The tech moves so quickly that you need to build towards what you think things will look like when you are actually ready to launch. With transaction preconfirmation on the horizon for Q1 2025, what other innovations can users and developers expect from Taiko in the near future? With transaction preconfirmation on the horizon for Q1 2025, what other innovations can users and developers expect from Taiko in the near future? With transaction preconfirmation on the horizon for Q1 2025, what other innovations can users and developers expect from Taiko in the near future? With transaction preconfirmation on the horizon for Q1 2025, what other innovations can users and developers expect from Taiko in the near future? The beginning of 2025 is all in on preconfirmations and eventually our DAO launch. These will further Taiko’s aims to become a more effective and decentralized rollup. From Gwyneth, we are working on building the development infrastructure. We’re excited to get some of the infrastructure up on testnet at the beginning of this year and launch it later in the year. It will be very busy, and we are excited to further discuss what is possible for scaling Ethereum. As Taiko prepares for its DAO launch, what steps are you taking to ensure community-driven governance aligns with the project’s ethos? As Taiko prepares for its DAO launch, what steps are you taking to ensure community-driven governance aligns with the project’s ethos? As Taiko prepares for its DAO launch, what steps are you taking to ensure community-driven governance aligns with the project’s ethos? As Taiko prepares for its DAO launch, what steps are you taking to ensure community-driven governance aligns with the project’s ethos? Ahead of the DAO launch, we will be introducing its architecture to our community to hear feedback and insights following the launch of the DAO on Helkla testnet. The build of our DAO is to ensure that Taiko is decentralized from the foundations all the way through - including how decisions are made.