On September 10, 2025 we launched the Third Cohort of the Decentralized Nodes program. A month later we look at what we achieved with this new cohort.
First off, though, the 3rd cohort was meant to launch on July 17, when the previous cohort ended. However, the plans to increase the size of the Polkadot active set necessitated that we rebalanced the stashes allocated to the program. For this reason we launched a “transitional” cohort, dubbed Cohort 2.1, with a selection of nodes participating in Cohort 2 for Polkadot and all Cohort 2 nodes for Kusama. Cohort 2.1 ran during the rebalancing process, and ended with the launch of Cohort 3.
For the third cohort we received 118 applications, mostly from returning but also new applicants. They were for 183 Polkadot nodes and 289 Kusama nodes. Even though these numbers were lower compared to Cohort 2, the quality of the applications was very high, making the selection process quite difficult. We had to select 99 Polkadot nodes and 194 Kusama nodes, which we did from 57 and 59 operators respectively. You can see all the selected nodes on the Decentralized Nodes site.
The application and selection process remained, to a very large degree, the same as in Cohort 2. We continued focusing on diversification to strengthen the decentralization and resilience of the networks. We favoured nodes that were hosted in remote and unsaturated locations, on less popular providers, or even better, were self-hosted by the operators, that featured Intel processors to improve the balance between the brands, and run on different Linux flavours. That being said, all applications were evaluated based on all the selection criteria and the final selection was based on the overall evaluation across all criteria. Special emphasis was given, as always, to the nodes’ specs, the operators’ technical expertise and their ecosystem affinity, while leaving room for new joiners.
And the results are quite satisfactory, as the following graphs show. For anyone wanting to compare, the Cohort 2 graphs can be found in this article. Obviously the more qualitative aspects of the evaluation, like the operators’ technical expertise or their involvement in the ecosystem, cannot be depicted in graphs, but we believe we’ve selected operators that are deeply involved and interested in the ecosystem, have the skills to provide high quality validation services, and the ethos to perform these duties diligently.
