Strengthening Cardano's foundations: Q3 2025 progress report

Summary:

• Q3 2025 saw steady progress across IOE projects, with a strong focus on scalability, assurance, and community collaboration
• UTXO-HD reached key integration milestones for efficient ledger storage, while Leios advanced through its first CIP and ecosystem prototypes
• Plutus Core v.1.50.0.0 delivered major tooling and performance updates, and the Cardano High Assurance team expanded formal verification with new Plinth features
• Hydra reached v.1.0.0 and refined its SDK roadmap, while the Acropolis modular node continued testing on SanchoNet
• Meanwhile, the KES agent audit began, nested transaction design matured, and revised stake pool incentive research progressed – marking another quarter of transparent, measurable innovation toward Cardano’s long-term roadmap.

The recent community vote in support of the Input | Output Engineering (IOE) funding proposal, under the Cardano vision and roadmap, represents more than just an allocation of resources. It reflects a collective commitment to advancing the platform through one of its most significant phases of technical development. With this funding, the focus now shifts to delivering on the roadmap: implementing performance-enhancing upgrades that broaden capabilities and solidify Cardano’s role as a sustainable foundation for global-scale applications. At the same time, delivery is guided by clear milestones, transparent reporting, and independent oversight to ensure accountability to the community at every step.

This Q3 report highlights the main achievements across several core and product workstreams, as well as progress in strategic initiatives.

Cardano core

UTXO-HD

The UTXO-HD workstream reached a key milestone in early Q3 with the successful integration of UTXO-HD on Lightning Memory-Mapped Database (LMDB). This approach enables Cardano to store ledger data efficiently on disk, rather than relying solely on memory, thereby improving scalability for larger ledgers. Internal performance benchmarking is now underway to compare disk-based LMDB with in-memory solutions.

In parallel, the team advanced integration of the Log-Structured Merge (LSM) tree library into the Cardano node. This work enables more efficient handling of ledger updates and snapshot conversions, reducing memory usage – replaying a chain required only 4 GB of RAM in recent mainnet tests. The team also added LSM support to the db-analyzer and finalized the pull request stack for full node integration pending the release of Cardano node v.10.6.0. 

Eidos – Cardano blueprint

The Eidos workstream remained on track throughout Q3, focusing on the development and evolution of Cardano blueprints. Efforts also included aligning project success criteria to ensure that the blueprint outputs support ongoing Cardano development.

Community engagement progressed through the collection of feedback, including surveys to assess sentiment regarding the blueprints. The team explored using artificial intelligence (AI) to enhance blueprint creation, investigating how a large language model could convert Haskell and Agda content into accessible markdown for broader developer use, with licenses procured and implementation plans underway. Product integration remained a priority, aiming to incorporate blueprints into future product requirement documents (PRDs) and streamline handover to relevant Cardano teams. 

Cardano products

Leios 24/7

Leios, Cardano’s major consensus upgrade, remained consistently on track and entered a decisive phase with the publication of its first Cardano Improvement Proposal (CIP) in August. The consensus, ledger, and cryptography teams completed or advanced impact analyses, while the Haskell-based simulator helped validate the performance of Linear Leios under various Plutus workloads.

Community and vendor engagement proved strong. TxPipe signed a statement of work for Leios support and review, while BlinkLabs and Serokell also contributed with analysis and feedback. At the same time, the development dashboard reached its first release, providing visibility into Leios metrics and APIs, and was presented at Token 2049.

The immediate priorities are to complete the impact analysis report, advance CIP reviews with partners, and translate the draft specification into actionable roadmap items. Strategic planning around Leios’s go-to-market approach, product objectives, and communications remains ongoing.

Plutus Core

The Plutus Core workstream remained on track throughout Q3, with the team focusing on core language enhancements, optimizations, and improvements to tooling. Key achievements include implementing casing support for booleans, integers, lists, pairs, and units, removing vestigial builtins, and progressing the `BuiltinValue` implementation. Performance improvements were delivered through cost model updates, specialized ArgStack optimizations, and preliminary benchmarking for array deserialization and BuiltinValue.

Tooling and infrastructure work advanced with end-to-end and conformance tests for new built-ins, bind tracing, and call trace features, as well as the launch of UPLC Comparative Artifact Performance Evaluation (CAPE) with defined benchmarking scenarios. Plutus version 1.50.0.0 was released, dependencies were updated, and the team explored AI/Plutus integration. Documentation and onboarding also progressed, including the development of new user guide resources, issue templates, and the addition of a full-time engineer. Additionally, the team initiated a statement of work for CIP-121 primitives that was scoped with MLabs.

Significant preparations were made for the UPLC 2025 workshop, covering participant registration, internal coordination, logistics with the University of Edinburgh, branding, website and landing page development, and the design of event materials. 

Cardano high assurance (CHA)

Work on Cardano High Assurance (CHA) progressed across multiple fronts. The Automated Formal Verification team advanced the formalization of ledger rules while delivering new features for the VS Code extension. Glyph (UPLC/RISC-V) expanded support for built-ins and made its repository public to increase accessibility. Other initiatives, including transaction monitoring and property-based testing, remained in the ideation phase, with vendor scoping underway to define future development paths. 

The team also delivered the proof-of-concept milestone for static analysis and built a static code analyzer in Plinth that implements four vulnerability checks focused on security and reliability. These checks have been validated on on-chain code, and the tool can be (and has been) integrated within Haskell Language Server (HLS) and merged into the main branch.

Acropolis

The Acropolis project, focused on building a modular node architecture, made notable progress between July and September. Reward state tracking moved from testing to implementation on SanchoNet. Work on the REST API has advanced, with a major refactoring completed and several new endpoints added, aligning the API more closely with the Blockfrost schema. Regarding infrastructure, the team addressed significant memory usage challenges and implemented performance instrumentation across all modules. Improvements such as the smarter Mithril Fetcher and pause-on-block support enhanced system flexibility.

Looking ahead, Acropolis will continue to refine reward calculations, expand Blockfrost endpoints, and complete API alignment. 

Hydra

Work on Hydra remained on track throughout the quarter, with the team prioritizing support for Midnight’s Glacier Drop. Hydra version 0.22.3 was released, upgraded across external Hydra head network operators, and later followed by 0.22.4, which introduced critical fixes for queue handling and message spam. Testing of Glacier Drop with Hydra integration was carried out in pre-production and on mainnet. The latest release v.1.0.0 is now out.

Beyond delivery support, the team strengthened its product focus. Efforts included backlog grooming, roadmap refinement, and early evaluation of Hydra SDK options with ecosystem partners.

KES agent

The KES agent workstream focused on integrating key evolving signature (KES) functionality into Cardano, a critical component for securing and signing blocks in a way that supports the Mithril protocol. Handling signing operations externally via a dedicated agent improves security and flexibility, allowing nodes to rotate keys safely without exposing long-term secrets. The work also includes coordinating a thorough security audit to ensure the implementation meets high-assurance standards.

Progress during the period included merging the KES agent into ouroboros-consensus, defining the Mithril interface for secure signing over sockets, and releasing an initial API version to CHaP. The team also finalized the audit contract with Bcryptic, conducted code walkthroughs, and began C bindings for BLS signatures, which are relevant to other Cardano initiatives like Leios. Testing and integration proceeded positively, though the final delivery faced delays due to external dependencies, including the late start of the audit and the pending cardano-node-10.6 release.

Nested transactions (Babel fees)

The nested transactions (Babel fees) workstream advanced its foundational research and documentation, keeping a ‘scoping’ status throughout the reporting period. A draft of the Babel fees PRD was developed, with version 1.0 expected for broader review by the wider ecosystem and collaborators in October 2025. The team also conducted a design analysis comparing smart contract- and ledger-based approaches, concluding that nested transactions offer the greatest value in open marketplace use cases where decentralized finance (DeFi) applications compete to fulfill user intents. 

External engagement expanded through meetings with TxPipe, Aquarium, Harmonic Labs, and the BitcoinDeFi/BitcoinVMX project, discussing potential applications and use cases for nested transactions.

Revised stake pool incentives

Data-driven analysis and stake pool operator (SPO) engagement formed the backbone of this workstream. The team developed queries and charts to better understand SPO distribution and trends while running focus groups with communities, such as the Japanese SPOs. A first milestone draft, identifying current issues with incentive schemes, is now in preparation.

Sustaining transparency and growth

Finally, the teams advanced audit and assurance by initiating the KES agent audit with BCryptic and beginning procurement for broader third-party assurance, including discussions with Dquadrant.

Overall, Q3 2025 marked steady progress across IOE’s treasury-funded workstreams. From advancing UTXO scalability and integrating new consensus mechanisms to formalizing incentives and building interoperability frameworks, the teams laid the groundwork for Cardano’s next phase of growth. These initiatives, together with oversight and independent audits, underscore the ongoing commitment to transparency, accountability, and community trust.