This dynamic and evolving technical roadmap, maintained and periodically updated by the Cartesi Foundation, summarizes Cartesi's ongoing efforts and technical direction.

A RISC-V virtual machine that boots a full Linux OS. Developers can use any language, library or tool that runs on Linux in a verifiable execution environment.
The PRT dispute system runs on Ethereum mainnet. A single honest party can enforce the correct result against a well-funded adversary.
Cartesi has achieved stage 2 on L2BEAT. One of a handful of rollups to reach the highest maturity classification.
Each application benefits from dedicated compute. Applications avoid block space competition while remaining deployable as an L2, an L3, or a sovereign rollup.
Last Updated on Aug 3, 2026
Financial applications need recovery paths and dispute resolution that hold under adversarial conditions.
Enable users to recover their funds if an application stops or a security council determines that funds are at risk, such as when an exploit is discovered).
Support multiple validator implementations, so that potential bugs in one implementation cannot be used to exploit the system.
Integrate dispute resolution across the full rollups stack, with hardened settlement and tournament logic
Subject the full solution to external and internal audits, extended live testing, and sophisticated attacks carried out with the help of AI agents
Real financial logic, like risk models, simulations, and statistical analysis, has to be cost-efficient enough to run every block.
Improve performance with faster state hashing, disk-backed machine state, more efficient proofs for long-running computations, and support for ZK proof.
Support ZK fraud proofs to eliminate delay attacks while still enabling arbitrarily complex computations beyond the capabilities of current ZK technologies.
Use a sequencer to provide fast soft confirmations for trading applications, with fraud-proof settlement behind them.
Linux onchain means developers can build with the languages, libraries, and tools they already know. Nothing new to learn.
Support multi-application throughput, restart resilience, production diagnostics.
Provide an MCP server and Cartesi skills so AI coding tools produce working Cartesi code.
Support local development, mainnet forking, node inspection.
Demos that include implementations covering risk engines, bonding curves, liquidity vaults, and advanced order types.
As the focus narrowed to DeFi, several other directions were discontinued: Bitcoin execution, the EigenLayer coprocessor, Lambada, alternative data availability integrations, sequencer integrations, ENS resolution, browser-based rollups.
The code remains open source, available to anyone in the community who wants to build on it or continue exploring these areas.
Follow Cartesi's ongoing technical evolution through GitHub releases, ecosystem updates and what is coming on our blog, and technical discussions on Discord.
© 2026 The Cartesi Foundation. All rights reserved.