Viral Wire

NVIDIA and Microsoft's RTX Spark Delivers 1 Petaflop AI — And Adobe Is Rearchitecting Photoshop for It

1 petaflop AI, 128GB memory, 6,144 cores — all in a Windows PC

Deep Dive

NVIDIA and Microsoft have unveiled the RTX Spark AI PC platform, a new Windows-based system designed specifically for running local AI agents, creative applications, and high-performance RTX graphics. The platform packs 1 petaflop of AI compute power — a figure usually reserved for data center GPUs — along with up to 128GB of unified memory, allowing large AI models and complex creative projects to run entirely on-device without cloud reliance. The underlying hardware features a Blackwell GPU configuration with up to 6,144 CUDA/RTX cores and a 20-core Arm-based CPU, aiming to balance raw AI throughput with energy efficiency.

Major OEM partners including Microsoft Surface, ASUS, Dell, HP, Lenovo, and MSI have already committed to producing RTX Spark-powered devices. Adobe is rearchitecting key creative tools like Photoshop and Premiere Pro to fully leverage the platform's AI capabilities, enabling real-time generative fill, intelligent video editing, and automated workflows that stay local. For developers, the platform supports popular AI frameworks and agent orchestration libraries, making it straightforward to deploy autonomous AI agents that operate securely on the user's machine. RTX Spark represents a significant shift toward democratizing high-end AI compute for professionals who need speed, privacy, and low latency without constant cloud connectivity.

Key Points
  • 1 petaflop AI performance on a local Windows PC, powered by a Blackwell GPU with 6,144 CUDA/RTX cores
  • Up to 128GB unified memory for running large models in-memory; 20-core Arm CPU for efficient task handling
  • Adobe is rearchitecting Photoshop and Premiere for the platform; OEMs include Microsoft, ASUS, Dell, HP, Lenovo, and MSI

Why It Matters

Brings data-center-grade AI to desktop for local agent deployment and creative workflows, cutting cloud costs and latency.

📬 Get the top 10 AI stories daily