I pitted the 32GB RTX Spark against a Snapdragon X2 laptop for local AI, and only one could actually run the model
NVIDIA wants the RTX Spark to be the laptop chip for running AI agents on your own machine, so I expected my $3,200 HP OmniBook Ultra 16 to run circles around a Snapdragon laptop. Instead, Hermes took one look at the 32GB RTX Spark and said every model I wanted was too big. The Zenbook A16, with 48GB of memory and Qualcommβs Snapdragon X2 Elite Extreme, loaded three of them in one click. It was slow, but for local AI agents, the cheaper laptop ended up actually getting the job done.
