Local AI Revolution: Nvidia's VP Hints at Shift to Home-Based Processing
Nvidia's VP of Product Adel el Hallak says 'local AI is here,' and it could change how people use AI at home. El Hallak has a DGX Spark, Nvidia's AI supercomputer, sitting in his house. He uses it to run AI jobs overnight, but what's interesting is that he feels guilty if he doesn't give it something to do before bed.
This behavior highlights the shift towards local AI, where processing happens on personal devices rather than in cloud data centers. Currently, most people interact with AI through browsers or apps, which means their data travels to a data center for processing and then returns. This approach has its drawbacks, including financial costs and privacy concerns.
Nvidia's DGX Spark is one of the first consumer-friendly local AI products. It retails for $4,699 and can run unlimited queries without per-token billing or usage caps. In comparison, a ChatGPT Plus subscription costs $20 per month, which translates to $1,200 over five years. This math suggests that owning a local AI device like the DGX Spark could be more cost-effective in the long run.
The ecosystem around Nvidia's local AI devices is becoming more consumer-friendly. Perplexity has launched Portable Computer, a system that packages a local AI model, agent tools, app connectors, and a sandboxed runtime into a single unit that runs on DGX Spark hardware. This setup allows users to run tasks locally before sending them to the cloud if needed.
Nvidia's RTX Spark laptops are another example of how local AI is becoming more accessible. These laptops will pack the same GB10 silicon as the DGX Spark and are expected to ship this fall from various manufacturers, including Dell, HP, Lenovo, Asus, MSI, Acer, and Microsoft.