Microsoft debuts Surface RTX Spark Dev Box to run large AI models without cloud costs
Summary
Microsoft unveiled the Surface RTX Spark Dev Box at Build 2026, a compact desktop with Nvidia's Blackwell-architecture RTX Spark processor and 128GB unified memory, delivering 1 petaflop of AI compute. Developers can run AI models exceeding 120 billion parameters locally without cloud API calls, challenging the per-token pricing model. Pavan Davuluri, Microsoft's EVP of Windows and Devices, noted such devices could run ~100B parameter models and emphasized large context (e.g., 100K tokens) requires significant memory, hence the 128GB unified memory pool. The device will launch later this year.
Why it matters
The product directly challenges the cloud-based per-token pricing model in AI, promoting local AI development, with significant implications for Nvidia, Microsoft, and the AI developer ecosystem.
Source links
Market reaction
The following is market reference information for related companies, and does not constitute investment advice.