TL;DR
-
Somnia opera como L1 Agentic, ejecutando agentes IA en cadena.
-
El token SOMI paga cómputo, participaciones y gobernanza de IA.
-
Mantiene compatibilidad EVM y añade validadores ejecutando modelos de IA.
Somnia pivots away from its earlier focus on metaverse and consumer applications and announces a repositioning as an agentic Layer 1, a chain that embeds artificial intelligence agents as the core of its infrastructure. The change rests on a complete overhaul of the leadership team. Peter Lipka takes the CEO role, joined by Harry Lang and Kevin Zia, while founder Paul Thomas concentrates on the long-term vision.
The network retains Ethereum Virtual Machine compatibility but adds an execution layer designed for smart contracts to consume real-time external data and run AI models directly on-chain. The immediate consequence for developers and institutional users is that automation no longer depends on external systems or traditional oracles. Decisions made by agents get validated by multiple verifiers before appearing on the block.
The platform processed more than 2 billion transactions since its mainnet launch in September 2025. The figure signals load absorption capacity. The new architecture uses the MultiStream consensus mechanism and the IceDB state database, two components the company employs to keep gas fees stable and reduce operational latency. That design matters for automated workflows where micro-transactions and frequent state updates follow one another. If fees do not remain low, the economics of agents become unworkable.
Agents Validated by Consensus, Not External Servers
Somnia introduces Somnia Agents, a solution that embeds AI computations inside the blockchain framework. A smart contract can launch a query to an external API, execute a model, and obtain a response that multiple validators check before accepting. No isolated node decides the outcome; consensus backs every step. Paul Thomas describes the result as a “market of markets,” because the capacity to react autonomously to changing data supports everything from sports predictions to betting systems inside gaming environments.
Reactive contracts do not wait for external instructions. They listen, process, and act according to the rules defined in their code. Decentralized finance and non-fungible tokens now sit in the background of Somnia’s roadmap. The company keeps them operational but lowers their priority in favor of autonomous agents.
The new executive leadership mirrors that pivot. Lipka brings scaling technology expertise, Lang covers market positioning, and Zia handles operational control. The profile combination aims to move the network from an experimental setting to a product capable of attracting institutional capital and developers needing real-time automation without off-chain dependencies.
The upcoming Prediction Conference 2026 in Las Vegas, cited by Somnia, will showcase use cases in gaming, insurance, and decentralized finance. Autonomous economic agents operating inside virtual worlds, content moderation systems, and AI-controlled non-player characters appear as the first beneficiaries of the new approach. The company claims that agent responses are cryptographically verifiable, eliminating dependence on external intermediaries and placing trust in the protocol.
The mainnet already demonstrated its processing capacity before the pivot. The 2 billion accumulated transactions now serve as a credential for institutional clients evaluating integration costs and operational stability. For treasury managers and market operators, Somnia’s proposal alters the calculation about where and how automation runs. Decentralized infrastructure reduces reliance on private servers but shifts risk to validator availability and the price of on-chain compute.
