Semi Doped Nvidia CES 2026
11 snips
Jan 12, 2026 The hosts dive into NVIDIA's CES 2026 keynote, highlighting the revolutionary DGX Spark, which brings server-class architecture to desktops. They explore the implications of on-prem AI and the historical evolution from mainframes to minicomputers. Discussions on the Vera Rubin platform showcase its integrated architecture and the rising importance of CPUs for agentic workloads. The Rubin GPU is designed for efficiency, with cutting-edge HBM4 technology. They also analyze market reactions to NVIDIA's announcements and competitive dynamics with AMD.
AI Snips
Chapters
Books
Transcript
Episode notes
Route Models Locally For Practical Agents
- Use a DGX Spark or similar on‑prem appliance as a local router for models and devices.
- Route visual queries to robots and text queries to fast text models to build practical agentic workflows.
DGX Station Is A 1,600W Beast
- Vikram Shekhar looked up DGX Station power and found it runs at roughly 1,600 watts, making it an enterprise-class device.
- He suggests DGX Station suits on‑prem inference serving rather than home hobbyists due to its power and cooling needs.
On‑Prem Appliances Expand Compute TAM
- Austin Lyons compares DGX Station to the minicomputer era that expanded compute TAM by making mainframe-class compute affordable.
- He argues on‑prem appliances convert some GPU OpEx into CapEx and unlock new local workflows for enterprises.
