- Cardano is planning a Layer 1 upgrade with Ouroboros Leios in 2026, which will increase throughput in the first mainnet version from around 10 to around 500 transactions per second (50x).
- Input Output is targeting a first public Leios testnet at the end of Q2.
Cardano is preparing a Layer 1 upgrade for 2026, which will increase transaction speeds by a factor of 50. At a community event in Tokyo on February 8, Input Output and Cardano founder Charles Hoskinson explained the plan for Ouroboros Leios, including a roadmap, target metrics and the ambition to deliver scaling without sacrificing decentralization and security.
Cardano will scale in 2026
Michael Smolenski, Product Manager for Cardano Core at Input Output, explained the initial situation:
“Our stake pool operators need to cover the cost of their operations from transaction fees rather than block rewards over the long term, and to do that they need network usage of around 50 transactions per second. So obviously we can’t stay where we are – at 10 transactions per second. We have to move forward.”
As a Layer 1 upgrade, Leios is intended to bring the throughput in the first mainnet version to 500 transactions per second: “With Leios we are achieving a 50-fold improvement in network capacity in the first release on the mainnet – we are going from 10 transactions per second to 500 transactions per second. But we are not stopping there. This is just the beginning.”
Smolenski announced a gradual scaling in order to slowly bring the infrastructure of the stake pools up to increasing requirements: a jump from 10-15 TPS to 10,000 TPS “must be done strategically”.
Instead of just emphasizing TPS, Smolenski also referred to a throughput measurement of the amount of data: the goal is “300 transaction kilobytes per second” in the first release, with a “confirmation time between 20 and 80 seconds”. The reason: Transactions and scripts have different sizes, which is why the data rate is more precise than a pure count.
From prototyping results, he deduced that at around 300 transaction KB/s, “confirmation takes between 20 and 50 seconds.” At higher speeds, the current tests show “traffic jams”; transactions would arrive later, end up in the mempool queue and therefore take longer. However, Smolenski emphasized that this is “just the first simple release” of Leios and further improvements are planned.
Smolenski named a concrete milestone in the roadmap: After research, planning, prototyping and simulations, they are “in active development” and are aiming for “the first public Leios testnet at the end of Q2”. There is no set date for the mainnet fork yet.
Hoskinson: Leios solves the trilemma
During his speech, Charles Hoskinson described Leios as the result of a long research:
“We have published more than two dozen papers and implemented dozens of protocols, and Leios is the product of more than 15 engineering firms over the years – and the product of 168 scientists over a decade, at institutions from Stanford to the Tokyo Institute of Technology.”
In terms of content, Hoskinson presented Leios not primarily as a TPS bet, but as a solution to the blockchain trilemma.
“Why Leios is special: It’s not about TPS. It’s solving the toughest problem in consensus and blockchain – the blockchain trilemma. We’re told you can only choose two: decentralization, security and scalability. This protocol is decentralized, secure and fast.”
Hoskinson also highlighted a special security mechanism that is intended as a throwback to the status quo:
“And the crazy thing is, if the protocol fails, then it fails to what we have today. It collapses back to the same protocol we use today.”
In this context, Hoskinson referred to November 21, 2025, when the Cardano blockchain experienced an unexpected fork due to a faulty transaction and resynchronized on its own: “Cardano split into two networks. And you know what it did? It came back together on its own.”
No Comments