Cardano’s Ouroboros Leios scaling architecture has demonstrated throughput of up to 1,000 transactions per second in large-scale simulations, providing an early benchmark for the consensus redesign now progressing through public testnet development. The 1,000 TPS result came from a simulated 1,000-node environment rather than Cardano mainnet or the current Leios testnet, making it evidence of technical potential rather than production capacity already available to users.
According to Input Output Research’s mid-year technology validation report, complex Leios variants using mainnet-like transactions reached 1,000 TPS, while simpler variants designed to require fewer infrastructure changes achieved around 100 TPS. The benchmark therefore represented several experimental protocol configurations, not a single finalized Leios design operating at a guaranteed 1,000 TPS. The project has since converged on Linear Leios as the architecture intended for Cardano mainnet.
Linear Leios Separates Ordering From Higher Throughput
Cardano’s current Praos architecture limits transaction inclusion to the capacity available when the elected slot leader produces a block, with blocks arriving roughly every 20 seconds. Linear Leios retains that underlying Praos chain while allowing producers to create much larger Endorser Blocks alongside normal Ranking Blocks. This parallel path lets additional transactions be processed and certified without replacing the security foundation provided by Praos.
The current design is also simpler than earlier research versions that relied on separate Input Blocks. Endorser Blocks reference additional transactions, while stake-weighted committees vote on them; once sufficient votes form a certificate, the endorsed transactions can enter the ledger through a later Ranking Block. Leios increases capacity by using bandwidth and computation that Praos leaves underutilized, rather than by sharding Cardano into independent transaction domains.
Operator requirements have also evolved as engineering moved beyond the original simulation phase. Current mainnet-readiness criteria target sustained performance on relatively modest hardware, including four CPU cores, SSD storage and around 10 Mb/s of bandwidth for the targeted 100–200 TPS range. Those engineering targets are more relevant to the present Linear Leios implementation than the earlier suggestion that operators would require six or more cores and 100 Mbps connectivity.
Public Testnet Replaces Simulation as the Main Benchmark
Leios entered a materially different phase in June 2026 when Musashi Dojo became its public testnet. The Earth phase subsequently ran for 41 days, produced more than 127,000 blocks and expanded from three participating pools to 63. Under synthetic load, that real network reached roughly six times the throughput Praos delivers on Cardano mainnet today, although the result is still a testnet measurement rather than evidence of production demand.
Development remains active rather than complete. In September, the consensus team was still refining transaction caching, Endorser Block download prioritization, memory limits and node behavior during synchronization. The current milestone is therefore mainnet readiness and adversarial resilience, not simply pushing a simulator toward a higher headline TPS number.
The sequencing of Cardano’s roadmap has also changed from the draft framing. The van Rossem Protocol Version 11 upgrade already activated on July 18, while Linear Leios is planned as a headline feature of the subsequent Dijkstra hard fork, targeted for late 2026 and still subject to on-chain approval. The next concrete milestone is Dijkstra’s progression toward mainnet and the phased activation of Leios parameters after sufficient stake pools register the BLS keys required for voting.
Jack Reynolds is SatoshiPick’s infrastructure mind. Based in Denmark, he follows Bitcoin, Ethereum, Layer 1 networks, stablecoins and DeFi with a practical question always in the background: does this actually make crypto work better?
His articles focus on the systems beneath the headlines: settlement rails, protocol upgrades, stablecoin usage, DeFi coordination and the regulatory limits around them. Jack avoids turning technical coverage into a maze. He explains what is live, what is still experimental and why the difference matters.