← Feed Deep Dive Matrix Subscribe

Perplexity’s local AI agent comes to Windows, but only for RTX GPUs with at least 24GB of VRAM

tomshardware.com 2026-09-15 Shane Downing
Entities
Tags
Local AIPerplexityNVIDIARTX GPUWindows SupportVRAM RequirementAI AgentIntelligent ComputingArtificial IntelligenceEdge AIDGX SparkMCP ServerCloud-Edge IntegrationAI ModelTask Automation
News Summary
Perplexity has partnered with NVIDIA to release a local AI agent for Windows, expanding functionality previously limited to Linux systems. The feature, named 'Portable Computer,' requires at least 24G... Read original →
Industry Analysis
Perplexity’s collaboration with NVIDIA to launch a local AI agent on Windows marks a significant push toward edge AI computing on consumer hardware. By setting a 24GB VRAM threshold, the move effectively standardizes GPU memory requirements across high-end cards like the RTX 3090/4090/5090, driving demand for upgraded memory configurations. This development accelerates the maturation of local AI toolchains and reinforces the shift toward secure, on-device processing, sidestepping cloud-based data transfer risks. In competitive terms, it intensifies pressure on Microsoft, Google, and others to enhance their own local AI offerings. Over the next 12–24 months, the rollout of RTX Spark PCs and similar devices will likely establish local AI computing as a core feature in consumer electronics, reshaping the ecosystem around GPU-centric edge intelligence.
Read Original Article →
Related
This page displays AI-generated summaries and metadata for research purposes. Original content belongs to the respective publishers.