Bloomberg Surveillance Nvidia Fails to Stoke AI Trade
11 snips
Feb 26, 2026 Drew Matus, MetLife market strategist who focuses on macro and Fed direction; John Bilton, JPMorgan multi-asset chief and former chemist who models asset allocation; Ted Mortonson, Baird tech analyst with semiconductor and AI infrastructure expertise. They debate Nvidia pricing and alternative silicon. They unpack how agentic AI reshapes compute, software economics, and portfolio positioning amid tech repricing.
AI Snips
Chapters
Transcript
Episode notes
From Naval Academy To Tech Analyst
- Ted Mortonson recounts surviving the Naval Academy's intense engineering program and 13 years flying maritime surveillance and anti-submarine missions.
- He links that disciplined engineering background to his current tech analysis approach at Baird.
Agentic AI Is Driving Massive Compute And Memory Demand
- The rise of agentic AI sharply increases compute and memory demand, putting NVIDIA at the center of the shift and creating shortages in memory capacity.
- Mortonson says agents let non-developers create code and spike consumption, intensifying training and inference needs.
Nvidia's Strategy Focuses On Share Through Efficiency Gains
- Nvidia aims for massive share across training and inference by reducing token costs and improving system design rather than simply price-gouging.
- Mortonson expects their Rubin cycle to lower per-token cost roughly tenfold and expand availability with a premium for functionality.