NVIDIA Personal AI Router, or PAIR, is a free open-source tool that finds compatible PCs on a home network and sends AI inference jobs to whichever machine has free capacity. Announced at IFA 2026 in Berlin, it forms the core of NVIDIA’s push to make local AI practical by tapping the computing power that sits unused in most households for large parts of the day.

More than half of U.S. households own two or more PCs. PAIR treats those machines as a shared pool instead of isolated devices. It works with the open-source tools Ollama and LM Studio and runs on Windows, macOS, and Linux. Compatible hardware includes GeForce RTX 20-series GPUs and newer, RTX PRO workstation GPUs, DGX Spark systems, and Apple M4 chips or later.

PAIR does not combine multiple GPUs into one bigger processor. It simply routes separate inference requests to different machines. That approach helps when AI agents break a complex job into parallel sub-tasks. In one NVIDIA test, five sub-agents running on a single device needed about 18 minutes. The same workload finished in under nine minutes when spread across three machines with PAIR.

One-Click Agents and Faster Inference

Setup friction has long limited local AI use. NVIDIA is addressing that with three agent applications that now install in one click on RTX and DGX systems: Hermes Agent from Nous Research, Perplexity’s Portable Computer, and OpenClaw. All three sit on llama.cpp and already include NVIDIA’s inference optimizations.

Performance numbers also improved. The llama.cpp backend reaches up to 1.9 times higher throughput on a GeForce RTX 5090. On two-DGX Spark clusters, vLLM delivers gains of up to 1.4 times.

RTX Spark PCs Ship in October

NVIDIA confirmed that its ARM-based RTX Spark Windows PCs will begin shipping in October under the product name N1X. Each system pairs a Blackwell-based RTX GPU with a 20-core Grace CPU, supports up to 128 GB of unified memory, and delivers up to one petaflop of AI compute.

At IFA, Lenovo showed the Yoga Pro 9n and Yoga 9n 2-in-1. Acer displayed a compact desktop design. Asus, Dell, HP, and MSI are preparing their own models. Game studios adding support include Electronic Arts, Embark, and Ubisoft, joining earlier partners Krafton, NetEase, Riot Games, and Xbox.

A set of models tuned for RTX hardware rounds out the local offering: Meta’s Muse Glimmer, DeepSeek v4 Flash, Qwen’s 3.8-Flash-Next, and NVIDIA’s Nemotron 3.5 Lightning.