IO & Midgard Labs: L2 Scalability Initiative

As a DApp developer, I want access to production-ready L2s that offer fast finality, high throughput, and sub-cent fees, so that I can build within the Cardano ecosystem and benefit from its superior security model rather than migrating to a competing ecosystem.

Opportunity: Cardano L1 delivers over two hours of finality, approximately $0.17 per transaction, and roughly 7-10 transactions per second (TPS). Competing platforms offer 400ms to 2s finality, sub-cent fees, and 1,000 to 7,000+ TPS. As a result, high-performance verticals such as decentralized finance (DeFi), AI agent micropayments, gaming, and consumer payments often exclude Cardano at the selection stage, even before a fuller technical evaluation begins. This prevents consideration of Cardano’s USPs: its superior security model, formal verification approach, and extended unspent transaction output (EUTXO) design.

Four structural gaps compound the problem:

• The Finality Gap: over two hours of L1 finality makes UX uncompetitive for DeFi, AI agents, gaming, and any other use case where transactions have to settle fast on-chain

• The Cost Gap: current L1 fees represent up to 1,000x overhead versus competitors, making high-transaction-volume applications economically unviable on L1

• The Developer Burden: the UTXO concurrency model introduces complex, centralized batcher infrastructure, stretching development timelines significantly compared to EVM chains

• The Throughput Ceiling: current L1 throughput of approximately 7-10 TPS constrains modern application networks that require far greater capacity

Upcoming L1 upgrades, including Leios (targeting approximately 400 TPS) and Peras (targeting approximately 2-minute finality), are expected to strengthen the base layer and expand its performance envelope. At the same time, high-performance use cases, such as zero-fee or sub-second interactions, can benefit from complementary scaling approaches. L2 solutions provide an additional path to address these requirements, supporting faster finality, lower costs, and higher throughput, while maintaining integration with the base layer.

L2 solutions extend the capabilities of L1 rather than compete with it. Use cases such as agent-based systems on Hydra or decentralized exchanges on Midgard illustrate scenarios that can benefit from dedicated execution environments.

By enabling these applications within the Cardano ecosystem, L2s can contribute to increased developer activity, ecosystem growth, total value locked (TVL), and L1 settlement usage. Activity on L2s ultimately anchors back to L1 through block publication and settlement, reinforcing the overall system efficiency.

Why Two L2s Are Required

Hydra and Midgard are not competing solutions. They are complementary layers that together address a broader range of high-performance use cases that Cardano currently does not fully support.

• Hydra is purpose-built for known-party, high-frequency environments. It requires a fixed, pre-agreed participant set where all parties remain online and co-sign state updates. This constraint is a feature: it is precisely what enables sub-second finality, privacy, and zero-fee transactions. An AI agent network on Masumi, in which a defined set of agents continuously exchange micropayments, is a perfect fit. A real-time gaming session between a known group of players, or a B2B payment rail between two enterprises settling at high volume, are use cases for Hydra. Hydra has already matured from research to market validation. The Midnight Glacier Drop successfully processed token distribution for 30+ million users across 8 blockchains. Delta DeFi is building a DEX that bets its entire product roadmap on Hydra. Masumi requires a machine-to-machine micropayment infrastructure that L1 can no longer sustain. These are not pilots; they are live production commitments.
• Midgard is purpose-built for open, permissionless environments where participants are anonymous, unpredictable, and do not run infrastructure. A public DEX or AMM, a lending protocol where borrowers and lenders are unknown to each other, an NFT marketplace, or any consumer-facing application with a large and changing user base requires Midgard. Midgard is Cardano's first truly permissionless optimistic rollup, made uniquely possible by Cardano's EUTXO architecture. Its UTXO-based design enables single-party fraud proofs (just one L1 transaction to prove invalidity and slash a malicious operator), eliminating the complex multi-round challenge-response games that make Ethereum L2 fraud proofs difficult to implement. Anyone can lock a bond in the on-chain Operator Directory to produce L2 blocks. Hundreds of user transactions are batched per block, driving costs below $0.01 per transaction, with costs decreasing as volume grows.

The distinction between the two solutions reflects different trust models rather than overlapping functionality. Supporting only one does not constitute a partial outcome; it leaves a meaningful coverage gap across use cases. With both systems in place, Cardano can address a broader spectrum of high-performance applications and better support builders across both open and restricted trust environments.

Co-Development with the Cardano Ecosystem

This proposal is structured as a deliberate co-venture between Input Output and Midgard Labs, the specialist team behind Cardano's first permissionless optimistic rollup. Midgard Labs brings a deep expertise in UTXO-based rollup architecture and its track record of advancing Cardano's L2 capabilities from concept to testnet.

A Portfolio Approach to L2 Scaling

This proposal represents the first step towards a comprehensive portfolio approach to L2 scaling for Cardano. The three-pillar structure reflects a deliberate co-investment across the full stack: shared infrastructure that benefits every current and future Cardano L2; Hydra hardening that protects the production use cases betting on Cardano right now; and Midgard's path to mainnet as the open-participation rollup the ecosystem needs.

The L2-agnostic infrastructure (data availability solution) benefits all current and future Cardano L2s. Investing in shared primitives prevents ecosystem fragmentation and solves technical bottlenecks once for the whole ecosystem rather than once per project.

Why now: Delta DeFi, Masumi, and Midgard are already building in the current cycle. The opportunity to strengthen Hydra into a platform that supports these teams is available today, but it is time-bound and may narrow as other ecosystems consolidate early traction. If these initial production use cases are not successfully supported, the impact would extend beyond near-term adoption metrics. It would also shape external perceptions of Cardano’s L2 trajectory, making future efforts to attract and support similar builders more demanding in terms of time and resources. Over time, this creates a compounding effect, where early execution choices influence the cost and difficulty of later recovery and expansion.

Proposed Value Delivered (Why)

Cardano's competitive position depends on whether it can offer competitive finality, fees, and throughput. L1 improvements (Leios, Peras) are coming but will not fully arrive until 2027 and by themselves, will not provide all the performance improvements required to make L1 competitive for high-performance applications. L2s are the only way to close the gap in the current cycle.

This proposal builds on years of prior ecosystem investment. Hydra represents significant research and development that has already been completed, and Midgard is approaching its testnet phase. Ensuring that early adopters of these systems are supported through hardening, infrastructure readiness, and interoperability is essential to realize the value of this work. Without this support, the opportunity to translate existing R&D into sustained production use may be reduced, requiring additional effort to re-establish confidence and rebuild momentum in subsequent cycles. The downstream effects include reduced developer retention, slower ecosystem growth, and diminished return on prior investment in Hydra research and development.

Beyond protecting existing investment, this proposal establishes a reinforcing cycle of adoption and development. Production deployments create reference points that reduce perceived risk for new projects, while demonstrated scalability helps attract builders who previously excluded Cardano due to early-stage performance constraints. As the L2 ecosystem matures, there is also a pathway for L2 teams to contribute back to Plutus and the L1, consistent with patterns observed in other ecosystems such as Optimism, Arbitrum, and Base, where L2 activity has supported improvements to core infrastructure. Midgard’s architecture also includes mechanisms that allow L2 operators to allocate a portion of sequencer revenue back to the Cardano treasury, similar in concept to how Optimism’s Superchain governance routes value from Base to the OP Collective. This establishes a structured, ongoing economic relationship between L2 activity and the Cardano treasury, aligned with long-term ecosystem sustainability rather than short-term extraction.

The Midgard Labs co-venture also directly contributes to Pillar 5 (Ecosystem Sustainability and Resilience) by distributing ownership of Cardano's L2 infrastructure across specialist teams rather than concentrating it within a single organization. This distributed stewardship model is the structural outcome that the Cardano 2030 Vision is designed to produce.

KPIs

KPI	Alignment	KPI Alignment Narrative
TVL	Yes: Fully	DApps on Hydra and Midgard lock funds on L1, directly increasing TVL. Open-participation Midgard unlocks public DEX and lending protocols, currently non-viable on L1. The global DeFi TVL opportunity in DEX and AMM alone is approximately $20 billion; Cardano's current share of approximately $150 million grows only with competitive L2 infrastructure. Delta DeFi and Masumi alone represent meaningful TVL growth if they scale on Cardano rather than migrate.
Monthly Transactions	Yes: Fully	L2s increase total transaction capacity dramatically. While some activity moves off-chain for settlement, net ecosystem transaction volume grows substantially as previously non-viable use cases (micropayments, high-frequency trading, gaming) become feasible. Hydra delivers sub-1-second finality and near-zero fees. Midgard targets 10,000+ TPS per instance with 1 to 2 second finality. Combined, these solutions expand Cardano's effective transaction capacity by orders of magnitude.
Monthly Active Users (MAU)	Yes: Fully	Finality and fee UX are the primary filters determining whether users acquire and retain. L2s bring Cardano to competitive parity on the two metrics that matter most at the acquisition stage. More viable DApps produce more MAUs. The pre-selection problem means that the current MAU figure understates Cardano's potential audience. L2s unlock the verticals that are currently self-eliminating.

Additional KPIs

KPI	Alignment	KPI Alignment Narrative
Reliability: Monthly Uptime (6 epochs)	Yes: Partially	Hydra’s operational excellence milestone directly improves node reliability. The program includes operator runbooks, observability tooling, and logging improvements to support high-uptime targets for production Hydra deployments.
Operational Resilience: Voting Power Distribution of Controlling Stake	N/A
Operational Resilience: Alternative Full Node Clients	N/A
Revenue / Adoption: Annual Protocol Revenue	Yes: Fully	Every Midgard block published to L1 generates L1 transaction fee revenue. Midgard's architecture is explicitly designed so that L2 activity flows back as L1 protocol revenue, a direct and ongoing contribution to treasury sustainability. Hydra lifecycle management similarly generates L1 fees. L2 fee accrual adds to the overall Cardano ecosystem revenue rather than allowing it to flow to competing ecosystems.
Governance: DRep Participation Rate	N/A
Scalability: Throughput Capacity per day	Yes: Fully	This is the primary KPI this program addresses. Hydra: sub-1-second finality, near-zero fees, unlimited parallel heads. Midgard: 10,000+ TPS per instance with 1 to 2 second finality. Combined, these solutions expand Cardano's effective throughput capacity by orders of magnitude for the verticals that need it most.

Pillars

KPI	Alignment	KPI Alignment Narrative
Pillar 1: Infrastructure and Research Excellence	Yes: Fully	The L2-agnostic infrastructure workstream directly strengthens Cardano's infrastructure foundation for all current and future L2 development. The data availability solution for every Cardano L2. Hydra and Midgard are both production L2 infrastructures for building high-performance DApps.
Pillar 2: Adoption and Utility	Yes: Fully	High-value verticals (DeFi, including DEX, AMM, perpetuals, and lending; AI agent micropayments; real-time gaming; consumer payments) currently avoid building on Cardano due to finality and fee constraints. L2s remove these disqualifiers. Development time for Cardano DApps drops significantly when sequencing infrastructure replaces bespoke batcher builds. Delta DeFi (DeFi vertical) and Masumi (AI agents and micropayments) are now critically dependent on Hydra, with more projects in the pipeline.
Pillar 3: Governance	N/A
Pillar 4: Community and Ecosystem Growth	Yes: Fully	Addresses the pre-selection problem: Cardano's current builder base has self-selected for tolerance of L1 constraints. L2s expand the addressable developer market to builders who require sub-second finality and sub-cent fees, providing the basic entry requirements for high-performance application development. The L2 Working Group creates active developer communities around each solution. The Midgard Labs co-venture introduces an additional specialist organization invested in Cardano's long-term success.
Pillar 5: Ecosystem Sustainability and Resilience	Yes: Fully	Midgard's L1 fee revenue model creates an ongoing, non-speculative revenue stream for the Cardano treasury from L2 activity. The L2 team flywheel (teams investing in Plutus and L1 as they invest in their own chains) strengthens the ecosystem's long-term sustainability. Decoupling transaction economics from ada price appreciation prevents the network from becoming too expensive to use during bull markets. The Midgard Labs co-venture distributes stewardship of Cardano's L2 infrastructure across specialist organizations, directly reducing single-point-of-failure risk and reflecting the distributed ownership model the Cardano 2030 Vision is designed to produce.

Deliverables and Roadmap

Sequence	Item Description
	Workstream 1 L2-Agnostic Infrastructure
Q3 2026	Cardano Data Availability (DA) Strategy & Prototype Stage 1: Evaluate and develop a number of candidate DAL designs that meet the decentralization, availability, security, tokenomics, and complexity requirements.
Q4 2026	Cardano Data Availability (DA) Strategy & Prototype Stage 2: Develop a specification and prototype implementation of the selected candidate design.
	Workstream 2 Hydra: Production Hardening + Turnkey Solutions
Q3 2026	DAG-Hydra feasibility study: Evaluate the applicability of DAG solutions from other chains to Cardano.
Q3–Q4 2026	Hydra reference implementations for DeFi: White-label-like solutions for DeFi use cases, including pre-built smart contracts, off-chain state machine, head lifecycle tooling, starter UI / API layer, and runbooks. Modeled on the learnings from Delta DeFi.
Q3–Q4 2026	Hydra Performance Optimization: Memory usage reduction, maximum TPS improvement, benchmarking suite, and binary protocol improvements.
Q3–Q4 2026	Hydra Operational Excellence: Resilience improvements, hydra-node operator recipes, easier node configuration, documentation improvements, improved TUI, observability and logging.
Continuous	Hydra Ecosystem Support: Specific feature requests from production users (Delta DeFi, Masumi, Midgard fast-withdrawal integration), hackathons and workshops, developer relations, Hydra Alliance / Working Group facilitation.
Continuous	Hydra Maintenance & DevX: CI improvements, tooling improvements, technical debt reduction.
	Workstream 3 Midgard: Optimistic Rollup to Mainnet
Q3–Q4 2026	Midgard Mainnet Launch: Penetration testing in preparation for mainnet release by end of 2026

Resources

Workstream 1: L2-Agnostic Infrastructure

• Innovation engineers (DA strategy Stage 1 and Stage 2)
• R&D engineer (DAG-Hydra feasibility study)

Workstream 2: Hydra Production Hardening and Turnkey Solutions

• Engineering from the Cardano ecosystem (Hydra reference implementations for DeFi)
• Hydra engineers across (performance optimization, operational excellence, ecosystem support for Delta DeFi, Masumi, Midgard integration)

Workstream 3: Midgard Optimistic Rollup to Mainnet

• Development engineer (Mainnet Launch 2026, Midgard Labs team)

Budget

Total Treasury Ask: ₳10,425,871

• Workstream 1 L2 Agnostic: ₳1,895,613
• Workstream 2 Hydra Production: ₳7,582,452
• Workstream 3 Midgard Optimisation: ₳947,806

Funding Distribution
Development	₳8,966,249	86%
Infrastructure	₳104,259	1%
Security & Audits	₳104,259	1%
Legal & Compliance	₳104,259	1%
Engagement & Ecosystem support	₳625,552	6%
Operations & Delivery	₳312,776	3%
Governance	₳104,259	1%
Others	₳104,259	1%

Pricing Principles: Midgard Labs and IO are requesting funding in ada and have provided USD figures as a reference. A portion of the funding shall be specifically tied to demonstrating measurable impact on Cardano's KPIs and pillars

• Personnel and Delivery: Majority of costs needed to fund the delivery resources across IO internal engineers, additional Cardano engineers & contractors, and Midgard Labs co-proposer contributions
• Infrastructure: Cumulative cost of machine instances for network, monitoring, testing, benchmarking, and testnet operations across both Hydra and Midgard workstreams
• Ecosystem support, Audit, Assurance & Contingency: Leadership, ecosystem, and delivery to support execution and wider alignment. Independent work assurance and audits, plus contingency to account for complexities during execution

Risks

Type	Description	Likelihood	Severity / Impact
Technical	Midgard's move from a single-sequencer testnet to a decentralized multi-operator set is the most complex engineering challenge in this proposal. Round-robin coordination introduces a new consensus surface area that has not yet been stress-tested.	High	High: This is on the critical path to mainnet. Mitigation: early testnet deployment, and oversight to identify blockers before they compound.
Technical	DA strategy selection requires choosing among multiple candidate architectures with materially different trade-offs. The wrong architectural choice creates long-term technical debt for all Cardano L2s that adopt the shared standard.	Medium	Medium: Stage 1 is an evaluation phase, not a commitment. The deliverable of Stage 1 is a recommendation with evidence; Stage 2 proceeds only with community input.
Community / Ecosystem	Delta DeFi and Masumi are live production users, depending on the completion of Hydra hardening on schedule. Delays in performance or memory optimization put these early adopters at risk of migrating to competing platforms, which would be a signal to the broader developer market.	Medium	High: This is the strongest external motivation for delivering WS2 on time. Mitigation: direct feedback loops with Delta DeFi and Masumi are built into the Ecosystem Support deliverable, with bi-weekly check-ins to surface blockers early.
Process	DAG-Hydra feasibility study (1-month sprint in Q3 2026) may conclude that DAG solutions from other chains are not applicable to Cardano's architecture. This does not block other workstreams but may require a scope adjustment within WS1.	Low	Low: The feasibility study is explicitly scoped as an investigation with two valid outcomes: adoption or documented rejection with rationale.
External	Hydra reference implementations for DeFi depend on 2 external contractors with deep Hydra knowledge. Finding and onboarding qualified contractors in Q3 2026 is a lead-time risk.	Medium	Medium: Contractor scoping and engagement to begin pre-approval. Delta DeFi's existing knowledge of Hydra makes them a natural candidate for a contractor or advisory engagement.

Additional Information

Known Limitations: The DA strategy deliverable is a specification and prototype, not a production deployment; mainnet deployment requires a subsequent proposal. Midgard multi-operator coordination is scoped for the testnet-to-mainnet transition; ongoing operator economics and GTM incentive design are dependencies that the proposal addresses at the framework level but not the full execution level.

Benefit Dependencies: Hydra hardening benefits depend on Delta DeFi and Masumi continuing to build on Cardano and acting as production validation partners. The decision on a dedicated L2 Scaling entity is under active exploration by IO and Midgard Labs and will affect long-term coordination structures beyond this proposal period.

Release and Solution Readiness: DA strategy specification and prototype: Q1 2027. Hydra reference implementations (perpetual DEX, prediction market, micropayment state machine): Q1 2027. Hydra performance and operational improvements: Q1 2027, with interim releases as milestones are achieved. Midgard multi-operator coordination: Q1 2027, enabling the path to mainnet in the subsequent cycle.

Treasury Governance & Compliance

Contract Management

A written off-chain Legal Contract will be created between Input Output and the Cardano Development Holdings (CDH), as mandated by the Constitution, and will be administered by Intersect. This will include details of the project delivery schedule and dispute resolution.

Project Delivery

All milestones, acceptance criteria, payment amounts and expected delivery dates will be agreed between the Input Output and Intersect, acting on behalf of the CDH. Input Output will deliver according to the agreed-upon project schedule within the Legal Contract, of which the necessary information will be made public via the budget management platform via transaction metadata.

Defined by the milestones within a Legal Contract, Input Output will submit and attest milestone acceptance to the community, Intersect or 3rd Party Assurer.

Project progress will be monitored via Intersect's delivery assurance function which will be communicated to the community.

Acceptance of the work will be supported by a 3rd Party Assurer, who will be responsible for reviewing and signing off the work completed at each project milestone against the corresponding milestone deliverables detailed within the Legal Contract. This work is funded from a portion of this treasury withdrawal.

Auditable Accounts & Fund Delegation

Budget Management Tooling

To administrate treasury funds on-chain, Intersect will utilize the treasury management smart contract framework developed by Sundae Labs. The smart contracts have been extensively tested, including audits from TxPipe and MLabs.

Final mainnet validation test can be seen via the Disburse action within transaction: 0f591dc544ae14102dbb4a74d5311a6acffc1772b163d8b7a9656b9525950b17

This withdrawal will utilize Intersect’s 2025 treasury reserve contract with the address being: stake17xzc8pt7fgf0lc0x7eq6z7z6puhsxmzktna7dluahrj6g6ghh5qjr
Funds will later be migrated to a 2026 treasury reserve contract once established.

Budget Management Specifics

Intersect will utilize a single Treasury Reserve Smart Contract (TRSC), with many Project-Specific Smart Contracts (PSSC), managed by Intersect. Intersect's management consists of three 'admin' and two Intersect 'leadership' roles. An Oversight Committee consisting of five external, independent third-party entities will provide checks and balances on Intersect, and safeguard against errors and unilateral control. The administration of both TRSC and PSSCs will be managed by Intersect, with external oversight on certain actions from the Oversight Committee.

The 2025 TRSC Oversight Committee consists of Sundae Labs, Cardano Foundation, Dquadrant, Xerberus and NMKR. Their role is to independently verify key administrative actions using on-chain logic, ensuring accuracy and consistency without exercising discretion over governance decisions.
For all details on Intersect's configuration please see the Smart Contract Guide on the knowledgebase.

The high level permissions are as follows:

• TRSC Fund and PSSC Modify
  • Two of the three Intersect admins, two of the five trusted entities and one of the two Intersect leadership sign-off must authorize
• TRSC Disperse
  • Two of three Intersect admins, three of five trusted entities and two of two Intersect leadership sign-off must authorize
• TRSC Pause and Resume
  • Two of three Intersect admins, and one of two Intersect leadership sign-off must authorize
• TRSC Sweep
  • One of three Intersect admins, and one of two Intersect leadership sign-off must authorize
• TRSC Reorganize
  • Two of three Intersect admins and three of five trusted entities must authorize

Processes

Upon enactment of this governance action, funding for this project will be directed into the TRSC's stake account. All instances of TRSC and PSSC can not be staked with a SPO and will be delegated to the auto-abstain predefined DRep. From here funds will be withdrawn into a UTxO remaining at the TRSC.

When a 2026 TRSC is established, the funding for this project will be migrated via the ‘disburse’ action.

When the Legal contract is prepared and Input Output is ready, funding for this project will be transferred using the Fund action to a PSSC. All milestones will be outlined within the metadata.

A dashboard will be available for the community to audit the TRSC or PSSC and track metrics related to this withdrawn ada as well as being immutably verifiable on chain.

Funding Denomination

All amounts in this proposal are denominated in ada (₳). The total Treasury ask is ₳10,425,871. USD figures ($2,502,209) are provided for reference only, based on an ADA/USD rate of 0.24.

Refund Conditions

All funds not disbursed by the end of the delivery period will be returned to the Cardano Treasury. A final reconciliation will be published as part of the oversight reporting cycle. In the event of partial delivery or scope reduction, unspent funds associated with cancelled or reduced deliverables will be returned proportionally.

Prior Treasury Receipts

IO and its affiliated entities has been accountable for delivery of work funded by the Cardano Treasury. The total funds allocated has been ₳130,708,860 across a number of projects within Treasury Smart Contract, to date IOG has withdrawn ₳78,459,777.

Workstream	Ada received	% of allocation	Corresponding Governance Action
Blockfrost	₳1,137,500	88%	8ad3d454f3496a35cb0d07b0fd32f687f66338b7d60e787fc0a22939e5d8833e#2
Catalyst	₳3,095,400	60% **	8ad3d454f3496a35cb0d07b0fd32f687f66338b7d60e787fc0a22939e5d8833e#23
IOE	₳47,159,487	49%	8ad3d454f3496a35cb0d07b0fd32f687f66338b7d60e787fc0a22939e5d8833e#1
IOR	₳26,840,000	100%	8ad3d454f3496a35cb0d07b0fd32f687f66338b7d60e787fc0a22939e5d8833e#32
Governance	₳227,390	38%	8ad3d454f3496a35cb0d07b0fd32f687f66338b7d60e787fc0a22939e5d8833e#22

**Note: for Catalyst this only reflects the workstream that focuses on the Hermes Infrastructure and UX/UI improvements, not the execution and operation of Funds 14-16. Per Info Action this is in the process of transitioning to Cardano Foundation.

Net Change Limit Compliance

The requested amount does not at time of submission, on its own or in aggregate, breach the applicable 350M Net Change Limit covering Epoch 613 to Epoch 713.

In accordance with the guardrail TREASURY-02a, this withdrawal does not exceed the NCL at the moment of submission.

Audit & Oversight

Audit and oversight costs are included within the overhead applied to this proposal. The Intersect administration fee covers administrative oversight and is reflected within the cost of this proposal. Independent oversight will be provided through Intersect and technically capable third-party, including reporting obligations and milestone-based disbursement controls.

Standardized Format & Immutable Hosting

Upon finalization, this proposal will be hosted on IPFS in an immutable format. The blake2b-256 hash of the document will be provided for on-chain reference and verification.