Fortytwo

Fortytwo operates a decentralized, peer-to-peer artificial intelligence network that utilizes swarms of small language models (SLMs) running on everyday consumer devices. This approach, termed "Swarm Inference," allows multiple specialized models to collaborate, cross-validate, and refine reasoning to handle complex tasks, positioning the company as an alternative to centralized AI giants like OpenAI and Anthropic. The network is designed to overcome the scalability and cost limitations of traditional AI by tapping into the idle computing resources of personal devices, creating a more cost-efficient, resilient, and scalable AI ecosystem that improves as more nodes join.

Founded in 2024 by CEO Ivan Nikitin, CTO Vladyslav Larin, and Chief Growth Officer Alex Firsov, the company is headquartered in Mountain View, California. The founders' background includes over a decade of developing AI projects, from self-learning agents to foundational models, where they consistently faced scalability challenges with centralized providers. This experience led them to develop Fortytwo as a solution for sustainable AI scalability. In March 2025, the company secured $2.3 million in a pre-seed funding round led by Big Brain Holdings, with participation from other firms including CMT Digital and Chorus One. This capital is intended to accelerate the rollout of its devnet and advance the development of its distributed multi-step reasoning capabilities.

The business model allows anyone to contribute their device's compute power by running a node on the network. In return for supporting inference tasks, these node operators earn rewards. For developers, Fortytwo offers access to AI reasoning through an API, enabling use cases such as coding assistants and research agents. The system's architecture is built to be open and permissionless, with a long-term vision of creating a community-driven, planetary-scale reasoning engine.

Keywords: decentralized AI, small language models, swarm intelligence, peer-to-peer network, AI inference, distributed computing, collaborative AI, swarm inference, AI scalability, node operators, consumer hardware AI, community-owned AI, reasoning engine, edge computing, decentralized protocol, multi-agent systems, AI infrastructure, blockchain AI, compute power sharing, distributed reasoning