At a glance
This proposal aims to make Cardano smart contracts:
- Cheaper to run, by adding new Plutus capabilities and built-in functions that reduce script size and execution cost.
- More expressive, by enabling contract patterns that are currently difficult, inefficient, or overly repetitive.
- More trustworthy, by strengthening formal specification, conformance testing, and structured review of critical evaluator and costing code. This is especially important in the era of node diversity and in the era of AI.
- Easier to build, by improving compiler and optimizer architecture, error messages, setup, and the overall user experience.
- More resilient as an ecosystem, by distributing Plutus stewardship across expert teams, with Input Output and VacuumLabs working together on core public infrastructure.
- The total treasury ask is ₳11,877,575 to fund delivery through Q2 2027, administered through Intersect's milestone-gated TRSC framework.
Where Plutus is now
Plutus is the compact, lambda-calculus-based language that Cardano nodes execute to validate transactions. Every smart contract on the chain, including DeFi protocols, zero-knowledge applications, and much more, runs through it. Its quality directly shapes what builders can do, how much it costs them, and how much confidence they can place in the result.
Today, Plutus provides the core foundation for Cardano smart contracts. But like all programming languages with real-world production use, it needs to keep evolving. We’ve identified several clear opportunities to make Plutus better.
Who is delivering this work?
Primarily the Plutus team at IO, which has been the main workforce behind the language and toolchain since its inception. The compiler, the CEK evaluator, the cost model, the metatheory, the Plinth surface language, and the Plutus Core specification all sit in this team's scope. The team has deep knowledge in Haskell, Agda, and compiler and runtime development.
Our treasury track record: ₳130.7M allocated across IO workstreams, ₳78.5M withdrawn, with IOR at 100% and Blockfrost at 88% of allocation disbursed.
Additionally, we'll be co-delivering with VacuumLabs, a specialist firm with experience in formal methods, security-critical Haskell, and Cardano infrastructure.
Three workstreams
We've scoped the work into three workstreams, each targeting a different part of the platform:
- Workstream 1: UPLC capabilities and primitives. This stream extends the UPLC execution model and the set of built-in functions. It delivers casing on the built-in Data type, which significantly reduces execution costs for most contracts because Data is everywhere in validators. It implements the already-ratified multiIndexArray builtin (CIP-0156) and the additional BuiltinValue functions defined in CIP-0168, turning common list and multi-asset operations into single, cheap built-ins. It investigates the redundant scope check and delivers either a CIP for its removal or a report explaining why removal is not advisable. It produces a CIP and benchmark-conditional C bindings for a SNARK-friendly Poseidon hash function. Finally, it opens the design space on laziness and memoization in UPLC with a CIP structured similarly to CIP-0152.
- Workstream 2: Formal specification, correctness, and security. This stream strengthens Plutus's foundations. It delivers a property-based conformance testing framework that extends the existing test suite with randomly generated programs, giving alternative node implementations a high-confidence way to check agreement with the canonical implementation across a wide program space. It funds a structured manual security audit of the Plutus evaluator and the costing logic, with findings delivered as a written report. And it formalizes the programmatic built-in types and functions in the Plutus metatheory in Agda.
- Workstream 3: Developer experience. This stream introduces a new compiler and optimizer architecture that improves code optimization, delivers clearer source-level error messages, and reduces boilerplate. It also simplifies setting up the development environment, removing the need for tools like Nix or manual dependency installation.
Learn more about the proposal by listening to this X Space.
What Cardano looks like when this is built
The target delivery date is the end of Q2 2027.
For users, the most visible change is in cost. Pattern-matching on Data, multi-asset comparisons, and multi-index lookups all become substantially less expensive. That makes smaller minimum trade sizes economical, makes more frequent automated actions like liquidations, rebalancing, and oracle updates viable, and makes Cardano a more competitive venue for capital-efficient protocols.
For builders, developer experience improves substantially. The development environment can be set up with a single command, without Nix or native libraries to install by hand. It is easier to write clean code without scattered boilerplate. Compiler error messages, which play a major role in user experience and retention, are significantly improved.
For the ecosystem, the correctness story matures. The property-based conformance framework gives alternative node implementers a serious way to check agreement with the canonical evaluator across a wide program space. A structured security audit of the evaluator and costing surfaces issues before they reach production rather than after. The metatheory becomes an authoritative specification for the built-ins it covers. The combined effect is a smart contract platform whose correctness guarantees keep pace with its growing surface area, even as node diversity expands. This will have a meaningful impact on TVL, since capital follows trust.
How the money is controlled
The ₳11,877,575 we've requested flows through Intersect's Treasury Reserve Smart Contract framework, with the Oversight Committee (Sundae Labs, Cardano Foundation, Dquadrant, Xerberus, NMKR) verifying key on-chain actions. Funds release against milestones, with independent third-party assurance before each payment. Undisbursed funds are returned to the Cardano treasury with public reconciliation. The two CIPs in scope, CIP-0156 and CIP-0168, were already ratified by the community; implementing them here is governance follow-through, not a new ask.
86% of the budget, ₳10,214,715, goes directly to development.
The decision before DReps
All programming languages used in real-world production need to keep evolving, and Plutus is no exception. The Plutus team has just completed development for the Van Rossem hard fork. The window between Van Rossem and Dijkstra is the natural place to land the next round of capability work, especially anything that requires a hard fork to activate. Also, the toolchain and onboarding improvements of the past year have changed the floor of what Cardano developers can expect, and continuing to invest now is what keeps the next wave of builders from bouncing off setup friction.
Besides, the cost of doing nothing rises with adoption. As more value sits on Cardano, and as more independent node implementations come online, the consequence of any correctness or security issue in Plutus is amplified. Investing in formal specification, cross-implementation conformance, and structured auditing now is significantly more cost-effective than responding to incidents in production.
Read the full proposal and cast your vote.
