Zylon is an on-premise enterprise AI platform purpose-built for regulated industries such as finance, healthcare, defense, and the public sector. It allows organizations to deploy generative AI entirely within their own infrastructure, whether on-prem servers, cloud VPCs, or fully air-gapped environments, with no data ever leaving the enterprise boundary. The platform is built by the founders of PrivateGPT, the open-source project with 57,000+ GitHub stars that has been used by organizations including Google, Meta, and J.P. Morgan. Zylon is the commercial, enterprise-grade evolution of that project, adding multi-user workspaces, governance controls, and production-ready deployment tooling. Zylon's stack has three main layers: AI Core (self-contained LLM inference, vector databases, and GPU orchestration), Workspace (an AI assistant and document creation interface backed by private knowledge bases and data integrations), and API Gateway (OpenAI-compatible endpoints with built-in authentication, rate limiting, logging, and observability for developers building on top of the platform). The company was founded in 2023 in Madrid, Spain by Ivan Martinez Toro and Daniel Gallego Vic. It raised a $3.2M pre-seed round in February 2024 led by Felicis, with participation from LifeX Ventures and Zypsy. As of early 2026 the team is approximately 15 people. In May 2025 Zylon announced a partnership with Telefonica Tech to expand private enterprise AI deployments. Zylon differentiates on fixed-cost, unlimited usage pricing with no per-token charges, and claims production-ready deployment in under one week via single-command setup. It targets enterprises that cannot use cloud AI services due to compliance, data residency, or sovereignty requirements. Key features: - Fully on-premise deployment (on-prem servers, cloud VPC, or air-gapped) - AI Core: local LLM inference, vector database, and GPU orchestration - Private Workspace: AI assistant, document creation, and knowledge base over internal data - OpenAI-compatible API Gateway with auth, rate limiting, and observability - No per-token pricing — fixed cost with unlimited usage - Single-command deployment, production-ready in under one week - Built on PrivateGPT open-source foundation (57k+ GitHub stars) - Designed for compliance-heavy sectors: finance, healthcare, defense, public sector
Not public (enterprise contracts; no per-token pricing model; fixed cost positioning)