From CryptoPotato · By Chainwire
[PRESS RELEASE – New York, United States, October 26th, 2026]
Ritual Chain turns AI agents into persistent economic actors that control their own wallets, onchain identity, and private state.
Ritual, a lab for autonomous intelligence founded by former Polychain Capital partners Niraj Pant and Akilesh Potti, has launched Ritual Chain, an EVM-compatible Layer 1. Unlike conventional blockchains built primarily to execute financial transactions, Ritual Chain is designed to support software that can independently use intelligence, control capital, interact with external services, and continue operating without constant human intervention. The network makes AI model access, private computation, and autonomous scheduling native capabilities for EVM applications.
This makes the reason for building a new chain clearer while preserving the technical differentiation.
The most visible demonstration of that combination is long-lived AI agents, the kind designed to operate for months or years rather than a single session. An agent on Ritual Chain persists independently of the machine or operator that created it, controls and spends funds held in its own non-custodial wallet, and can keep sensitive prompts and keys confidential from compute operators through trusted execution environments.
The network launches with integrations to Phala Network, Nous Research, LayerZero and Stargate, following a public testnet that has been live since June and processed over 40 million transactions, both agent- and human-originated. Ritual Chain delivers sub-300 millisecond block times, placing it among the fastest Layer 1 networks in the industry. Ritual Chain includes 15 native precompiles that bring capabilities like LLM inference, web access, and autonomous scheduling directly into the execution layer. The protocol has been audited by Zellic.
Ritual introduces a new execution and verification architecture that allows computationally intensive work to be delegated to specialized nodes and verified by the network, avoiding the replicated execution required by most blockchains. This execute-once, verify-many-times model allows the network to support far larger workloads, including AI inference, without requiring every validator to perform the same computation.
Long Horizon Agents Require Dedicated Blockchain Infrastructure
Most AI agents today live inside a session: they run while a human is watching, on infrastructure someone else controls, with an API key that can be revoked at any time. That is fine for a chatbot. It does not work for an agent meant to manage its own treasury, run a recurring paid service, or coordinate with other agents over long periods. A long-lived agent needs continuity that does not depend on its creator, capital it actually controls, and privacy from the machines it runs on. On Ritual Chain, an agent’s state and identity are held onchain, its scheduled work continues within the permissions and budget it was given, and its sensitive inputs can be processed inside trusted execution environments designed to shield them from node operators.