Nvidia announced in September 2026 the release of Personal AI Router (PAIR), a free, open-source software tool that connects multiple home computers into a unified system for running local AI inference tasks.
Despite its name, PAIR is not a hardware router. The software discovers compatible devices on a home network, links them together, and coordinates them to handle agentic AI workloads using tools such as Ollama and LM Studio. Compatible hardware includes Nvidia GeForce RTX 20-series GPUs and newer, RTX Pro GPUs, DGX Spark systems, and Apple M4 chips or newer. The beta is available now for Windows, Linux, and macOS.
PAIR is designed to use participating machines only when they are idle, so it does not interfere with other tasks. If a user starts playing a game on a desktop PC, for example, PAIR can adapt and redistribute the workload. The system allows multiple devices to process requests in parallel, which Nvidia says reduces bottlenecks that would otherwise occur on a single GPU.
At a media briefing, Nvidia product manager Seth Schneider illustrated a household scenario featuring an RTX Spark laptop, a DGX Spark desktop, an RTX 5090 laptop, a gaming desktop, and a MacBook Pro — estimating the combined compute at roughly 165 teraflops. Schneider described this as “a treasure trove of free tokens just sitting in homes today.” He said Nvidia expects most users will have a more modest setup, such as one laptop and one gaming PC.
Security is handled through an initial six-digit pairing code, with ongoing communication encrypted via mTLS (Mutual Transport Layer Security), creating a trusted, encrypted channel between devices.
Alongside PAIR, Nvidia announced that three AI agent applications — Perplexity Portable Computer, Hermes Agent, and OpenClaw — will offer simplified local setup on Windows with Nvidia GPUs, reducing the manual configuration previously required to get local agents running.
Source: The Verge