Perplexity’s on-device AI offering promises data control and lower token costs
Perplexity on Tuesday rolled out an offering that runs the AI entirely on a local machine, and that, it said, will keep “private data local and escalating to the cloud only when a task needs it.” The service, called simply Portable Computer, is a local version of Perplexity Computer that runs on the Nvidia DGX Spark with Qwen 3.8 27B or with PPLX 27B, a post-trained version of the Qwen model,…
Perplexity has introduced a new service called Portable Computer that runs AI entirely on a local machine. This service keeps data private by processing it locally and only transmitting it to the cloud when necessary. The software runs on an Nvidia DGX Spark with Qwen 3.8 27B or PPLX 27B, and a 30B open model is expected soon. It requires Linux as its operating system, with Windows support coming soon.
The benefits for enterprises include data remaining local and not consuming token credits, as work isn't performed in the cloud unless explicitly instructed. This trend is growing, as users aim to reduce token costs and minimize reliance on expensive high-end AI systems. Portable Computer allows users to escalate tasks to the cloud for current information, browser use, connected apps, or advanced reasoning models.
However, some consultants and analysts question the economics, hardware requirements, and control over cloud usage. They argue that the hardware demands are steep, and the lack of clear details on local-versus-cloud decision-making raises concerns. Users might inadvertently grant cloud access, and there's a risk of unauthorized cloud escalation.
Enterprise administrators would need to provide a mandatory egress proxy, DLP inspection, and immutable logging to ensure comprehensive data egress governance.
Written by urgent.news from Computerworld's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
