Tezos X: From Roadmap to Reality
The Tezos X roadmap is about to become reality. First up: EVM and Michelson apps sharing one ledger.
9 minute read
This is a joint post from Nomadic Labs, Trilitech, and Functori.
When first introduced in 2024, the ‘Tezos X’ roadmap described an ambitious vision for Tezos:
A next-generation blockchain architecture with multiple tool stacks running as a single system, starting with EVM and Michelson. High performance, full composability, and broad interoperability.
Fast forward to 2026: the prerequisites are in place, priorities have been sharpened by market signals, and we are happy to announce that the architecture enabling EVM and Michelson to run together as one system is about to ship – starting with a testnet in April.
This blog post covers:
A recap of Tezos X
What’s coming (and when)
Milestones reached
How priorities changed
Tezos X, a recap #
Tezos X is a new execution layer for Tezos. It’s where transactions run and smart contracts live – designed to remove friction for users and expand what builders can ship.
For users, Tezos X offers experiences that simply work. Imagine an artist listing an NFT priced in tez on a marketplace written in Michelson (Tezos’ native smart-contract language). A buyer wants to pay in USDC, held by an EVM smart contract. On Tezos X, both contracts share the same chain and the same ledger, so the swap and the purchase settle atomically in a single transaction. No bridge, no wrapped tokens.
For builders, Tezos X offers new designs and frictionless development. A Michelson contract can call an EVM contract (and vice versa) inside one transaction, letting a single app tap user bases and liquidity that used to be siloed. Contracts deployed on the Michelson runtime can also be formally verified – mathematically proven to behave as specified – for extra assurance in audit and in production. EVM developers bring existing Solidity contracts and tooling directly. Michelson developers keep working as they do today – no rewrite needed.
Under the hood, Tezos X is an enshrined, non-custodial rollup: a fast execution layer that settles back to Tezos Layer 1 for security, built into the Tezos protocol itself, with users always in control of their assets and free to exit.
What’s coming now #
The execution layer will initially offer two interfaces:
EVM (Etherlink)
Michelson (available on the Tezlink Shadownet testnet)
It will be introduced as an Etherlink upgrade proposal that adds a Michelson interface, effectively evolving Etherlink into the execution layer.
This approach makes the execution layer instantly EVM-compatible, while the Michelson interface brings compatibility with Tezos’ Layer 1. The architecture enables additional interfaces in the future (for example JavaScript), and more will come.
The planned launch of Tezos X was also covered by Tezos co-founder Arthur Breitman in a keynote at the TezDev conference in March 2026. It’s recommended viewing for a high-level update on current protocol and ecosystem developments.
A frictionless path for existing Tezos apps #
The Michelson interface offers a way for existing Tezos applications built on Layer 1 to get the benefit of Tezos X.
The vast majority of Michelson contracts can be redeployed as-is and work out of the box.
Developers can keep using the wallets, connectivity libraries, explorers, and other tools they already rely on, thanks to each interface’s interoperability with its origin ecosystem.
Etherlink tools for Layer 1 interoperability (bridges, fast withdrawals) will be extended to include the Michelson interface.
Layer 1 itself continues to evolve toward a lean, fast consensus layer, supporting the existing ecosystem and with XTZ and FA tokens easily transferable to and from the execution layer.
The timeline #
The expected near-term timeline is the following:
May 2026: Testing. A new testnet launches with Michelson and EVM interfaces on a single shared ledger. This is the first chance to test native atomic cross-interface calls between Michelson and EVM contracts in practice. Developers will be encouraged to deploy, test, and break things.
June 2026: Etherlink governance vote. After testnet validation, an Etherlink upgrade proposal is submitted to bakers. If approved, Tezos X becomes a reality on mainnet, starting with Michelson and EVM interfaces – live and fully composable.
H2 2026: RISC-V migration. The rollup engine migrates from WASM to RISC-V, enabling more runtimes, JIT compilation, and more predictable gas accounting. Critically, it makes adding new interfaces significantly faster, opening the door to mainstream programming languages later.
Milestones reached #
The launch of Tezos X on mainnet builds on numerous improvements and innovations introduced since 2024.
Layer 1 has been continuously optimized for speed, efficiency, security, and decentralization. Block time has been reduced to now 6 seconds, while the staking UX and economics have been improved and fine-tuned. In short, Layer 1 is rapidly becoming a fast, lean consensus layer for Tezos X.
The Data Availability Layer (DAL) has been launched, putting a check mark next to a major milestone on the roadmap. The DAL ensures that Tezos has the bandwidth required for publishing millions of transactions per second, while security and integrity remains guaranteed by Layer 1. The bandwidth is continuously being improved – an upgrade to 10 MB/s is part of the upcoming Ushuaia protocol proposal.
The launch of Etherlink was another key milestone. Besides the added EVM interface for Tezos, Etherlink, Smart Rollup technology offers execution scalability far beyond what can be achieved on Layer 1, while still being non-custodial and governed by Tezos bakers. Etherlink has in many ways been a prototype for Tezos X, with much of the innovation implemented here first.
Priorities: what moved up, what moved down #
Though the roadmap and vision have been broadly stable since 2024, priorities have been adjusted in response to market signals.
Some work has moved up on the agenda – and some things have been added – due to necessity or demand, while other items remain on the agenda but are no longer part of the near-term path.
Note that any property mentioned for Etherlink will carry over into the unified execution layer, should the upgrade proposal be adopted.
What moved up
Throughput: During 2025, Etherlink’s maximal throughput was increased almost 14x to 27 Mgas/s (~1300 TPS), with experiments showing further potential for significant increases.
Latency: Work on latency has already enabled 10–20x faster confirmations than originally envisioned, from “subsecond” to milliseconds. Etherlink now offers instant confirmations via a sequencer, enabling 50ms latency today and further reductions in the future.
MEV protection: The instant confirmations provided by the sequencer also offer guarantees about transaction order and the result of the next block, effectively providing MEV protection.
BLS signatures: This signature scheme was introduced to enable signature aggregation, supporting Layer 1 optimization and DAL integration, and bringing protocol-native multisig functionality.
What moved down (for now)
JavaScript interface: The absence of strong demand at this stage means that enabling a JavaScript interface (Jstz) in Mainnet is no longer a near-term priority. An important milestone, the RISC-V migration—required to support additional runtimes such as JavaScript—is targeted for H2 2026, as mentioned above.
Blind sequencer: The current sequencer is decentrally governed by Tezos bakers, holds sequencing power only, and can be replaced by vote. On-chain settlement adds censorship-resistant finality in seconds. Our market analysis indicates that, at this stage, the ultra-low latency achievable with this approach is a stronger value driver for network participants than alternative architectures with higher latency.
Data Availability Sampling: Sampling enables higher scalability by establishing security about available data without requiring an honest majority of nodes to download everything. However, at 10 MB/s bandwidth, the DAL supports plenty of throughput for the near future, and sampling is therefore not a near-term priority.
Finally, an initially envisioned automatic upgrade of Layer 1 apps to the execution layer has been dropped. Instead, app maintainers can move ad hoc, supported by tooling to make the process smooth.
From “canonical rollup” to “execution layer” #
Some readers may have noticed that “execution layer” has replaced previous talk of a “canonical rollup”.
While “canonical rollup” partly captures the role Tezos X has in the Tezos architecture in a technical sense, the term “canonical” is not an intuitive descriptor for most people. Also, “rollup” carries Ethereum L2 connotations that don’t match what Tezos X represents:
Ethereum’s L2 landscape is fragmented into largely independent and custodial chains with varying security assumptions, offering limited integration with Ethereum’s Layer 1 and each other.
Tezos X is a unified, single-chain experience for both developers and end users, enshrined in the protocol. No new tokens or controlling entities are involved. The system remains entirely non-custodial.
We believe that “execution layer” better reflects the tightly integrated nature of the architecture, and it sums up the role well: the natural home for applications on Tezos.
Beyond blocks and chains #
If the Tezos X roadmap laid out an ambitious path for Tezos, this update marks the point where that path becomes reality.
It is an opportunity for developers and entrepreneurs to think bigger, draw on the best of the EVM and Michelson worlds, and deploy the next generation of Tezos-powered applications and products.
We look forward to supporting builders in exploring the new possibilities through documentation, tutorials, tools, sparring, and other help that makes the path to great products as frictionless as possible.
The long-term belief hasn’t changed. Blockchains should fade into the background. People should use applications that happen to be powered by Tezos, not “use a blockchain” as a primary act. A seamless experience resting on uncompromising security.
Tezos X is what makes it practical: a fast, multi-language, highly composable execution layer secured by a battle-tested, self-amending, and censorship-resistant Layer 1.