BREAKING
NVIDIA Runs AI Agents Live on DGX Spark
0
GB
unified memory
0
PFLOP
FP4 AI perf
0
$
price
Model Sizes on One Unit
Inference
200
Fine-tune
70
Autonomous x86-to-Arm Migration
1
Legacy x86 app
↓
2
Arm MCP Server
↓
3
Ported to Arm
↓
4
Runs locally
0
x
inference speedup
0
day
AI Enterprise trial
A Base for Always-On Agents
AI NEWS BLITZ
NVIDIA shows autonomous AI agents migrating and running directly on its DGX Spark desktop supercomputer.