Microsoft debuts Surface RTX Spark Dev Box to run large AI models without cloud costs
VentureBeat / 6/3/2026
📰 NewsDeveloper Stack & InfrastructureSignals & Early TrendsIndustry & Market MovesModels & Research
Key Points
- Microsoft unveiled the Surface RTX Spark Dev Box, a compact desktop designed to run large AI models locally to avoid cloud computing costs.
- The device uses Nvidia’s Blackwell-architecture RTX Spark processor and includes 128GB unified memory, enabling developers to run models with over 120B parameters without making cloud API calls.
- Microsoft says model effectiveness depends not only on size but also on sufficient context length; at 100,000 tokens of context, key-value cache can require 40–50GB of memory.
- The Dev Box will be available later this year in the United States exclusively via Microsoft.com, with pricing not yet disclosed.
- By positioning fixed, upfront hardware costs against per-token or metered cloud pricing, Microsoft aims to challenge the dominant AI economics since ChatGPT’s launch.
Continue reading this article on the original site.
Read original →Related Articles
Black Hat USA
AI Business
Microsoft launches MXC, an OS-level sandbox for AI agents, with OpenAI and Nvidia already on board
VentureBeat
Microsoft created the mini Surface dev box that Qualcomm couldn’t
The Verge
Trump signs narrower executive order on AI oversight after industry objections
TechCrunch
Naive RAG vs Agentic RAG: The Evolution of Intelligent Retrieval
Dev.to
