IOTA's Bid for the Next DAG-Based Breakthrough

Key Takeaways

• Since launching its mainnet in 2016, IOTA had consistently pursued a vision of building a feeless value exchange network based on its unique DAG-based consensus mechanism.

• To support a wider range of use cases and drive broader adoption, the IOTA team recognized the need to introduce smart contracts—enhancing the flexibility of asset representation and enabling more diverse functionality.

• Based on this understanding, the team initiated the development of the IOTA EVM Layer 2 in 2022 and has since been advancing a strategic overhaul of its Layer 1 protocol under the banner of “IOTA Rebased,” incorporating the latest technical stack inspired by the Sui blockchain.

• In parallel, IOTA continues to develop its own consensus engine and roll out initiatives aimed at improving both user and developer experience, positioning itself as another differentiated DAG-based smart contract platform.

1. Introduction

It’s already been a decade since the first smart contract platform emerged, and in that time, the blockchain industry has advanced at an astonishing pace. One key factor behind this rapid growth is the open-source nature of blockchain ecosystems, which gives them both forkability and composability. Made possible by the ethos of transparency at the heart of blockchain, these traits have played a crucial role in helping countless builders flexibly strengthen their business logic or reduce the switching costs between different platforms as they work to design better services and infrastructure.

Despite this environment, integrating something you did not build yourself into your existing (or planned) service is no easy task—especially for services that have already absorbed significant time and resources. Not only must the original value the service provides remain consistent, but in an ecosystem where countless projects are all experimenting with different ideas, there is also a need to carve out and maintain a unique identity. In this context, it’s especially noteworthy that IOTA, which debuted in 2015 as a leading Directed Acyclic Graph (DAG)–based network, has recently undergone a significant evolution — embracing a MOVE-based architecture through the integration of Sui’s high-performance infrastructure to unlock a new era of scalability, programmability, and ecosystem growth.

Launched in late 2015, IOTA’s core vision can be summarized as “Built to Make a Difference.” From the outset, IOTA has pursued a feeless value exchange network based on its own consensus mechanism, all in service of realizing that vision. But in striving to meet its goals, it has encountered numerous technical challenges. Each time, the IOTA team has responded by repeatedly refining its infrastructure through mathematical theory and simulation—and, when necessary, by boldly discarding old infrastructure in favor of new concepts.

A consistent vision, strong in-house research and engineering capabilities, and a willingness to embrace multiple rebrandings in the name of improvement: these are the hallmarks of the IOTA team. So why did they choose to carry out a large-scale rebasing onto Sui MOVE, and what new directions do they plan to explore through this transformation?

Deep in-house research and engineering expertise as well as the agility to adapt to evolving market demands — these are the hallmarks of the IOTA team and the foundation of its long-term resilience and innovation. So why did they choose to carry out a large-scale rebasing onto MOVE on Sui, and what new directions do they plan to explore through this transformation?

2. Consistent Vision, Repeated Improvements

2.1 Overview of DAG

For several years, IOTA has been striving to create an infrastructure optimized for token transfers. As part of that effort, it chose to employ a Directed Acyclic Graph (DAG). Widely used in fields such as network theory and topology, the concept of a DAG attracted significant attention at the time as a structural alternative to existing blockchain infrastructures, which suffer from low scalability and high fees. Many projects began replacing the block structure itself with a DAG and developing new consensus algorithms to support it. This movement even led some to label the trend as “Blockchain 3.0.”

The key difference between a traditional blockchain and a DAG-based structure lies in the consensus process. Rather than having all validators synchronize and agree on a single block creation at the same time, each validator signs and verifies transactions individually, based on its own view of the network state. If a transaction manages to obtain the signatures of a supermajority of validators within the network, that transaction reaches consensus and becomes final. By taking the aggregate of all such confirmed transactions, the overall state of the network can then be inferred. Thanks to this approach, a DAG can handle transactions asynchronously and in parallel, enabling a more flexible and highly scalable system than traditional blockchains.

Over time, however, DAG-based blockchains began revealing structural limitations of their own—challenges in achieving interoperability and overly centralized node operations, for instance. This led to fewer projects successfully reaching commercial viability, and public interest gradually waned.

Even so, research into a new paradigm—where only the consensus mechanism is DAG-based rather than the entire blockchain structure—has continued. Indeed, the Sui network has adopted a DAG approach in its network architecture, and Aptos has indicated that it may also introduce DAG-based consensus to improve scalability. As a result, DAG is receiving renewed attention.

2.2 Improvement, Improvement, and More Improvement: IOTA’s Challenging Journey

In its early days, IOTA adopted a DAG-based ledger called the Tangle - the Tangle did away with conventional “blocks” and generalized a single-chain structure into an expanded graph by having each new transaction reference two previous blocks. Under this system, a consensus on transaction validity could be reached without involving separate miners, which made the feeless nature of the Tangle one of its most distinctive features. The figure below offers a simplified illustration of the Tangle’s structure.

Because multiple transactions could occur simultaneously on the Tangle, the network could exhibit the unique characteristic of handling greater throughput as it grows. However, maintaining this structure while ensuring stable and decentralized consensus required a special mechanism. To address this, IOTA initially introduced a node called the "Coordinator," which periodically issued checkpoints. While this approach enhanced both security and structural stability, it inevitably came at the cost of sacrificing some degree of decentralization.

Ultimately, in order for IOTA to realize its overarching vision of “Digital Autonomy,” it was essential to ensure a consensus in which all nodes participate equally. To that end, the IOTA team carried out extensive research and continuous experimentation on distributed consensus algorithms such as “Chrysalis (IOTA 1.5)” and “Coordicide (IOTA 2.0).” The insights gained through these efforts laid a critical foundation for IOTA’s continued evolution, culminating in the strategic shift announced in 2024, and labeled as IOTA Rebased.

2.3 And, IOTA Rebased

IOTA Rebased represents a groundbreaking evolution, where IOTA leverages the cutting-edge technology stack of Sui and redefines its protocol for enhanced scalability, speed, and efficiency.

Before adopting Sui’s technology, the IOTA team had been following a roadmap (IOTA 2.0, Coordicide) optimized for token transfers. However, they believed that to support more diverse use cases and drive wider adoption, it would be necessary to enable the flexible representation of assets and a broad range of functionalities through smart contracts. Described as “L1 programmability,” this initiative spurred the IOTA team in 2022 to form an internal group focused on evaluating both the technological feasibility and the overall solution environment for integrating smart contracts into IOTA’s Layer 1 along with the development of the IOTA EVM Layer 2*.

The IOTA team concluded that, among all available solutions, the MOVE Virtual Machine (MOVE VM) developed by Mysten Labs—the creators of Sui—was the most applicable and promising option for integration with the IOTA’s DAG-based ledger, particularly in terms of security and long-term potential. Sui is a Layer 1 blockchain that operates with MOVE-based smart contracts in the form of an object-oriented model. Since its mainnet launch in 2023, it has garnered significant attention in the industry for its high-performance parallel execution.

Over several months, the IOTA team modified Sui’s Rust codebase and integrated IOTA’s unique design philosophy. Throughout this process, the core concepts that IOTA had been pursuing—parallel execution, a DAG-based mempool and consensus mechanism, and proof-of-stake consensus—were faithfully reflected on top of the Sui’s codebase.

The result of these efforts is IOTA’s new Layer 1 protocol - IOTA Rebased. Completed in late 2024, it is built upon Sui’s codebase, whose stability and performance have already been proven. As a result, IOTA Rebased naturally brings MOVE-based native smart contracts to the IOTA network.

The outcome of these efforts is the new IOTA Layer 1 protocol, IOTA Rebased, which was finalized in late 2024. It is built on the proven stability and performance of Sui's codebase, enabling IOTA Rebased to seamlessly integrate MOVE-based native smart contracts into the IOTA network.

*To support Solidity-based smart contracts, IOTA continues to offer the IOTA EVM as a separate Layer 2 solution.

3. Beyond Token Transfers, Toward L1 Programmability

3.1 Like SUI

After navigating multiple twists and turns over the years, IOTA Rebased now aims to combine the strengths of DAG and blockchain technologies. It aspires not only to provide fast transaction speeds and high flexibility but also to operate as a more decentralized smart contract platform that guarantees a wide range of functionalities.

Currently, IOTA Rebased leverages Sui’s Mysticeti consensus algorithm to achieve low-latency transaction sequencing and finality. By adopting the MOVE programming language, it enables safer and more cost-effective asset definitions while supporting smart contract implementations that align seamlessly with the underlying network structure—such as parallel transaction processing.

3.1.1 The MOVE Language for Flexible Smart Contract Definitions

As mentioned earlier, a core objective of IOTA’s rebasing is to evolve into a smart contract platform capable of flexibly defining various assets and managing their interactions. This calls for a programming environment in which diverse functionalities can be developed and supported in an extensible manner.

However, Ethereum’s existing EVM and Solidity stack come with structural limitations that make truly seamless implementations challenging. For instance, the EVM environment stores the same logic repeatedly (e.g., duplicating contract deployments), and it relies on a single-threaded model that hinders parallel processing. Moreover, each smart contract must be trusted independently, and because standards that are difficult to implement at the protocol level have mostly been introduced in a fragmented manner at the application layer, chronic issues such as scalability, processing efficiency, and memory stability become even more entrenched.

By contrast, the MOVE language—developed by former Meta engineers (including Mysten Labs co-founder Sam Blackshear)—was designed to address these fundamental bottlenecks, security vulnerabilities, and inconsistencies in token standards. Instead of relying on a global state structure, MOVE separates state on a per-object basis, allowing asset relationships to be defined more explicitly and enabling smoother parallel execution of logics. This design achieves significantly higher TPS (transactions per second) on the same hardware while greatly reducing fees. Further, MOVE can treat resources as first-class values that cannot be duplicated, discarded, or altered at will. Its data abstraction capabilities also conceal complex resource management and security logic within modules, simplifying both development and usage.

MOVE further bolsters security through static verification, which flags potential errors and vulnerabilities related to types, expressions, functions, and variable usage before the code is executed. A borrow checker prevents unintended data corruption or broken references, while formal verification can be used to mathematically prove the correctness of a smart contract’s logic.

IOTA has optimized these MOVE features for its own infrastructure. By adopting an object-based architecture, it significantly improves parallel execution and scalability. At the same time, the use of module initializers and flexible entry-point designs ensures that even complex business logic can be implemented securely. In this way, IOTA is creating an environment that is both developer-friendly and highly secure, meeting the need for flexible asset representation and extensive functionality support, and laying a solid foundation to fulfill its vision as a next-generation smart contract platform.

3.1.2 Mysticeti for the High-Performance Consensus Engine Optimized for DAG

To ensure the smooth processing of a wide range of assets and contracts defined in MOVE, IOTA sought to optimize the transaction flow in its DAG-based network. In doing so, it adopted Sui’s “Mysticeti” consensus engine. Mysticeti is a DAG-based consensus protocol newly proposed by Mysten Labs to reduce the latency of Sui’s original Bullshark engine*. Designed to avoid reliance on a single leader and make full use of the network’s bandwidth, Mysticeti enables multiple validator nodes to process transactions in parallel while still arriving at a rapid total ordering.

*Bullshark is a round-based asynchronous DAG consensus engine, designed to flexibly assign anchors (block leaders) according to a threshold parameter for block commitment. This approach aims to help validators achieve an agreed ordering with minimal communication overhead.

Source: MYSTICETI: Reaching the Latency Limits with Uncertified DAGs

Mysticeti significantly reduces latency by enhancing the Bullshark engine through three key innovations. First, it eliminates the traditional certification process that required collecting signatures from a quorum of validators in every round (i.e., uncertified DAG), and instead introduces implicit certificates, allowing blocks to be finalized with as few as three message exchanges. While Bullshark waited for multiple validator signatures before broadcasting a block to the network, Mysticeti streamlines communication by having each validator broadcast its block immediately after signing it.

Second, it simplifies the consensus process by removing the primary-worker structure that existed in Bullshark. While the Bullshark engine required a two-step procedure—where worker nodes gathered transactions and a primary node proposed blocks—Mysticeti eliminates this waiting period by allowing transactions to be included directly in blocks.

Third, multiple leader blocks are selected in each round, enabling more transactions to be finalized immediately. In the Bullshark engine, only one leader block could be finalized at a time, which meant that some transactions included just before might have their finalization delayed until the next round. Mysticeti resolves this issue by allowing multiple blocks to be designated as leaders simultaneously.

Consequently, by integrating the Mysticeti consensus engine, IOTA preserves its original DAG-based transaction flow while achieving the high performance, robust network security, and decentralized consensus mechanism exemplified by the Sui network—laying a solid technical foundation for supporting increasingly complex user interactions.

3.2 Uniquely IOTA

In addition, IOTA has independently pursued a range of initiatives aimed at further enhancing the network’s operational efficiency and developer-friendliness.

3.2.1 The Starfish Consensus Engine for Greater Stability and Scalability

Released as open source on April 1, Starfish is a new consensus protocol developed by IOTA to enhance the stability and scalability of the previously adopted Mysticeti engine. While Mysticeti focuses on low latency and a simplified architecture, Starfish aims to achieve both security and scalability through a partially synchronous DAG-based BFT approach.

As mentioned earlier, Mysticeti’s uncertified DAG structure is effective for achieving low latency. However, because it omits the certification step that was required each round in prior consensus protocols, it inevitably sacrifices some degree of security compared to fully certified approaches that require every communication round. Consequently, Starfish—positioned as IOTA’s next-generation engine—proposes a middle ground: it preserves much of the efficiency of an uncertified DAG while incorporating robust security features similar to those of a certified DAG. More specifically, Starfish is designed in a way that allows the availability of the transaction data in a block to be proven after a certain period has passed since a block was proposed, aiming to minimize the possibility of malicious nodes omitting or concealing parts of the data.

This goal is accomplished via a technique called Encoded Cordial Dissemination, which combines Reed-Solomon erasure coding with data availability certificates (DAC). In essence, when a block is broadcast, encoded shards are pushed out to the entire network, and the leader block serves as a certificate confirming data availability for past blocks. In this approach, each node fully propagates its own block, but disseminate only one encoded shard of blocks from other nodes. As long as f+1 shards are verified through Merkle proofs, the original transaction data van be recomstructed by Reed-Solomon decoding. Thanks to this structure, data can be disseminated efficiently and without loss—even in large-scale networks or in the presence of Byzantine nodes—and it maintains high throughput under heavy transaction loads.

Additionally, Mysticeti’s pull-based data retrieval approach requires each node to request the data it needs, which can increase duplicate requests and network latency if the network is unstable or blocks are generated very quickly. While Mysticeti is designed to commit multiple leader blocks per round for faster consensus, any delays in data propagation can lead to repeated follow-up requests, creating potential bottlenecks. In contrast, Starfish validates data availability ahead of the consensus process through the leader block. Although this requires one extra communication round, it provides a higher level of stability in large-scale transaction environments.

In short, both protocols aim for secure consensus that tolerates Byzantine faults. However, while Mysticeti emphasizes speed and simplicity, Starfish places greater weight on stability and scalability.

3.2.2 Gas Station for a Seamless User Experience

In addition to its consensus engine, IOTA has developed various building blocks to enhance the experience of both developers and end users. One of these is the Gas Station.

The IOTA Gas Station is a tool designed to let developers and application providers sponsor network transaction fees on behalf of their customers, thereby simplifying user onboarding and lowering barriers to Web3 adoption. With this, users do not need to hold IOTA tokens or pay fees themselves when interacting with applications. Meanwhile, developers can take on transaction costs through the Gas Station and leverage features like access control and customizable limits to optimize gas management more flexibly. In other words, the Gas Station delivers a straightforward and intuitive user experience while streamlining the complexities of fee management for applications.

3.2.3 IOTA Indexer for Smooth Data Retrieval

IOTA also operates a node data service that aggregates and provides both original data and derived data generated on the IOTA chain. Checkpoint blobs gathered from the full nodes are indexed in a relational database and visualized in various table structures. This setup enables developers to run a reader binary in the form of a JSON RPC server, allowing them to handle online transaction processing (OLTP) requests efficiently.

Next to the JSON RPC API, IOTA also provides a GraphQL API for running powerful queries on the indexed chain state. This flexible framework simplifies data queries for developers and lets them select only the data they need in responses, making the overall process more efficient.

Developers can also take advantage of the open-source custom indexer framework to index and expose application specific data in their dApps. With such detailed insights into on-chain state, events and user interactions, graphical user interfaces can be made responsive and informative, all elevating the dApp experience.

3.2.4 The IOTA Identity Framework

IOTA envisions a future in which people, organizations, and even devices can all seamlessly interact without additional trust with digital identities and verifiable credentials. To this end, it has built the IOTA Identity Framework, which serves as a universal trust layer for the internet, delivering secure and transparent digital identity management across a wide range of sectors.

With this framework, individuals can obtain self-sovereign identities (SSI) that let them choose which information to disclose, based on each use case. Organizations adopting this solution can comply with necessary regulations while using data in a cost-effective, privacy-enhancing manner. As for devices, they can be assigned unique global IDs that allow for the identification and verification of their attributes—making fully realized digitalization possible.

3.3 Tokenomics & Roadmap

Alongside a major overhaul of its technology stack and protocol-level achievements, IOTA has also revamped its tokenomics - over the past decade, as the network’s consensus mechanisms underwent multiple iterations, the utility and overall tokenomics of IOTA’s native token, $IOTA, likewise evolved.

In the original IOTA Crowdsale in 2016, 100% of the original IOTA supply of appr 2.779 billion $IOTA tokens was sold for 1337 BTC (around 500K USD at that time) to the early IOTA Community*. However, given that the Tangle framework at the time did not require miners or stakers—and hence imposed no transaction fees—the token had little intrinsic utility besides being a medium of exchange.

Because $IOTA had a fixed supply yet lacked a strong use case, its tokenomics underwent its first major reconfiguration in October 2023, coinciding with the introduction of the IOTA Stardust network. The primary goal of this overhaul was to foster ecosystem growth. Through a governance vote, the total supply was expected to expand from 2.779 billion to 4.6 billion tokens, with roughly 1.82 billion newly issued tokens allocated to various areas - of this new supply is subject to a vesting schedule, releases gradually over two to four years with an initial unlock of 10%. Key allocations included 7.1% for the IOTA Foundation (R&D), 12% for the Tangle Ecosystem Association (TEA), 12% for the UAE-based IOTA DLT Foundation, 5% for rewarding past contributors and partners, and 3.5% for an airdrop to Assembly network stakers.

*As 100% of the supply was sold via the IOTA crowd sale in 2016, the community donated 5% of the supply to fund the creation and operation of the IOTA Foundation as a German non-profit Foundation in 2017.

More recently, with the launch of IOTA Rebased, the consensus engine has transitioned to a Delegated Proof-of-Stake (DPoS) model that introduces a deflationary mechanism by burning transaction fees to offset inflation. Under this approach, the original distribution plan remains intact. However, by removing the supply cap and introducing a fee-burning mechanism, the system strengthens network security while incorporating deflationary pressure—ultimately establishing a more flexible and sustainable economic model that prevents uncontrolled token inflation.

As IOTA develops new infrastructure solutions for smart contracts, real-world use cases for the $IOTA token are expected to multiply. Accordingly, factors such as base gas fees, optional validator tips, storage deposit costs, and staking rewards (about 767,000 IOTA per epoch, approximately 6% annual inflation) will manage the token’s supply and demand - The staking rewards are freshly minted in every epoch and distributed to the validators staking pool at the end of every epoch providing ongoing incentives for network participants.

At the time of writing, around 3.78 billion tokens are in circulation, according to Tokenomist. Going forward, IOTA plans to further refine its tokenomics—focusing on continued consensus improvements and potential local fee market implementations.

4. Looking Ahead

A DAG-based BFT consensus approach separates data propagation from consensus and allows each node to interpret the DAG locally. As a result, even if some nodes are slow or malicious, the others can continue finalizing blocks without interruption. This flexibility not only addresses the consensus delays caused by leader dependence in conventional BFT protocols like Tendermint and HotStuff, but also markedly boosts performance and scalability, all while preserving decentralization and strong security.

Against this backdrop, it is especially noteworthy that IOTA—having long used DAG as the backbone of its value-transfer architecture and successfully applied it, supported by extensive research, across sectors such as finance, supply chain, IoT, gaming, healthcare, social media, and the public sector—has decided to fork Sui’s technology as a technological baseline to accommodate an even broader range of use cases, making a full-scale transition into a smart contract platform. Moreover, the team is now moving forward with plans to introduce its own consensus engine, aiming to build a more distinctive infrastructure.

If IOTA’s unique research depth and fast-paced engineering execution—combined with its plans to support a diverse range of virtual machines beyond MoveVM and EVM, protocol-level initiatives tailored for developers, and the operation of its Business Innovation Program to drive ecosystem growth—can come together into a coherent and sustained trajectory, then there is ample reason to expect that IOTA could emerge as another meaningful success story among DAG-based networks, following in the footsteps of Sui.