Your idle PCs just became a local AI cluster
Nvidia has released Personal AI Router (PAIR), a free, open-source software tool that discovers compatible computers on a home network and links them together for local AI inference. Despite the name, PAIR is not a physical router. It's software that lets machines running tools like Ollama and LM Studio share work across a disaggregated set of home devices.
PAIR is designed around idle resources. When a desktop or laptop isn't actively being used, PAIR can assign AI workloads to it, leaving the machine free to return to normal use without interference. The system works in parallel, which is especially useful for agentic workflows that break complex tasks into smaller sub-tasks. Nvidia says PAIR adapts dynamically as devices join or leave the network—if someone starts a game on a desktop, PAIR routes work elsewhere.
Compatibility spans a broad range of hardware:
- Nvidia GeForce RTX 20-series and newer GPUs
- RTX Pro GPUs and DGX Spark systems
- Apple M4 chips or newer on MacBooks
In a media briefing, Nvidia product manager Seth Schneider described an extreme example: a household with an RTX Spark laptop, a DGX Spark desktop, an RTX 5090 laptop, a gaming desktop, and a MacBook Pro could have around 165 teraflops of underutilized compute. "It's truly a treasure trove of free tokens just sitting in homes today," he said, even accounting for electricity costs.
Security and realistic deployment
PAIR pairs devices using a six-digit code, then secures the connection with mTLS (Mutual Transport Layer Security). This creates an encrypted, mutually trusted channel between computers, so local AI traffic isn't exposed on the network.
The PAIR beta is available today for Windows, Linux, and macOS. When asked who PAIR is really for, Schneider said Nvidia expects most users to have a modest setup—typically one MacBook or Windows laptop plus one gaming PC. That's a far cry from the extreme household example, but still enough to offload substantial AI work.
One-click local agent setup
Alongside PAIR, Nvidia announced simplified local setup for three major AI agent apps on Nvidia GPUs with Windows:
- Perplexity Portable Computer
- Hermes Agent
- OpenClaw
These new setup experiences aim to get local agents running in a few clicks, reducing the manual configuration typically required. For users who already own capable hardware, PAIR lowers the barrier to running private, local AI without paying per-token cloud costs.