NVIDIA Levels Up Local AI Agents Across RTX PCs and DGX Spark

2026-06-01 · Source: NVIDIA Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Cloud Computing & IT Infrastructure · Depth: Intermediate, medium

Summary

NVIDIA introduced RTX Spark, a new class of Windows PCs designed for personal AI agents, at GTC Taipei at COMPUTEX. These PCs boast 1 petaflop of AI compute and 128GB of unified memory to support secure, private on-device agent execution. Concurrently, NVIDIA launched DGX Station for Windows, an AI deskside supercomputer for professionals. The company is collaborating with Microsoft to establish a robust Windows platform for agents, integrating new security primitives and the NVIDIA OpenShell runtime for user control and privacy. Key performance enhancements include 2x inference speed for agentic models via multi-token prediction in llama.cpp and vLLM, alongside multi-GPU optimizations for llama.cpp and ComfyUI. Adobe is rearchitecting Photoshop and Premiere for RTX Spark, promising up to 2x faster AI and editing. Additional updates encompass NVIDIA Broadcast 2.2, Blender Cycles integrating DLSS 4.5 Ray Reconstruction, and RTX Video Frame Generation. The DGX Spark OS for Linux also received updates for streamlined agent deployment and faster inference.

Key takeaway

For AI Engineers and creative professionals considering local AI agent deployment or performance optimization, NVIDIA's new RTX Spark PCs and DGX Station for Windows offer purpose-built hardware and software. You should evaluate these platforms for secure, private on-device agent execution, leveraging features like OpenShell and multi-GPU optimizations. This enables significantly faster inference for agentic models and enhanced creative workflows in applications like Adobe Photoshop and Premiere, transforming your local AI capabilities.

Key insights

NVIDIA is driving secure, high-performance local AI agent adoption across its RTX and DGX hardware and software platforms.

Principles

• Secure, private local agents demand dedicated hardware and software.
• Multi-token prediction significantly boosts agentic model inference.
• Multi-GPU optimizations enhance memory and compute for AI workloads.

Method

NVIDIA OpenShell, with Windows security primitives, provides identity, containment, policy, and privacy features for on-device agents, including intelligent query routing and personal data disguise for cloud interactions.

In practice

• Deploy local agents on RTX Spark PCs for 1 petaflop AI compute.
• Apply multi-token prediction in llama.cpp for 2x inference speed.
• Configure multi-GPU setups for llama.cpp (tensor parallelism) or ComfyUI.

Topics

• NVIDIA RTX Spark
• AI Agents
• Local AI Inference
• Multi-GPU Optimization
• Windows AI Platform
• Creative AI Applications

Best for: AI Architect, Computer Vision Engineer, AI Product Manager, AI Engineer, Machine Learning Engineer, AI Hardware Engineer

Related on AIssential

• Nvidia stretches compute power to Windows PCs in support of agents — Nvidia is expanding its full-stack AI compute strategy by bringing powerful agentic AI capabilities directly to Windows PCs and deskside supercomputers.
• NVIDIA Partners With Microsoft on Unified Stack for Agentic AI Deployment, From Windows Devices to Cloud to Local — NVIDIA and Microsoft are building a full-stack, secure, and scalable ecosystem for agentic AI, from edge devices to cloud.
• Nvidia’s RTX Spark Laptops Look Hell-Bent on Disruption — Nvidia's RTX Spark laptops, with integrated N1 CPUs and unified memory, aim to enable powerful local AI processing on Windows.
• Now Generally Available, NVIDIA RTX PRO 5000 72GB Blackwell GPU Expands Memory Options for Desktop Agentic AI — The NVIDIA RTX PRO 5000 72GB Blackwell GPU expands desktop AI capabilities with increased memory for larger models.
• Open-Source AI Tool Upgrades Speed Up LLM and Diffusion Models on NVIDIA RTX PCs — NVIDIA's CES 2026 announcements significantly boost AI PC developer capabilities through framework optimizations and new model releases.
• New Software and Model Optimizations Supercharge NVIDIA DGX Spark — NVIDIA's DGX Spark leverages NVFP4 and software optimizations for significant local AI model performance and memory efficiency.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by NVIDIA Blog.