Perplexity Launches Portable Computer for Linux RTX GPUs With 24GB VRAM, Removing Cloud Dependency

Linux systems equipped with NVIDIA RTX GPUs and at least 24GB of VRAM can now run Portable Computer, a fully local version of Perplexity Computer. The software executes the orchestrator LLM, subagent LLM, and agent harness entirely on user hardware to remove cloud dependency. A compatible version for Windows systems will follow the current Linux release.

The rollout includes a launch of Portable Computer on NVIDIA DGX Spark, which enables users to perform agentic tasks without sending data to external servers. This architecture differs from traditional AI apps by keeping all processed data and model execution on the device. CEO Arav Srinivas stated that the Linux launch expands the accessibility of its on-device AI ecosystem.