Euangelion original explanatory figure
Created: 2026-11-09
Creator: Euangelion / Codex
Identity: Authored vector explanation; not a documentary photograph or product screenshot.
Rights: Original SVG composition and text prepared for Euangelion; no third-party visual asset reused. Source facts are attributed below.
Substantive meaning: Conceptual planned inference path within an MGX rack: prompt processing, or prefill, on an NVIDIA GPU; handoff through NVLink Fusion; response token generation, or decode, on a Rebellions R200 NPU. This is a schematic of the announced division of work, not a literal final hardware layout. R200 systems are expected to be ready for deployment in 2028; no throughput or energy improvement has been measured in the announcement.
Limits: Planned GPU prefill and NPU decode share a rack-level architecture; readiness is expected in 2028.
Source basis:
https://nvidianews.nvidia.com/news/nvidia-rebellions-and-sk-telecom-announce-collaboration-to-advance-sovereign-ai-infrastructure-with-nvidia-nvlink-fusion
https://www.globenewswire.com/news-release/2026/11/09/3398037/0/en/nvidia-rebellions-and-sk-telecom-announce-collaboration-to-advance-sovereign-ai-infrastructure-with-nvidia-nvlink-fusion.html
https://news.sktelecom.com/232944
https://www.benzinga.com/markets/tech/26/11/63030222/nvidia-expands-south-korea-ai-push-with-new-startup-investments
Final desktop/mobile browser inspection remains the publication coordinator's responsibility.