Neil Movva Discusses AI Inference on Podcast
Podcast episode features Neil Movva discussing AI inference layers, chips, power, and spending.
Patrick O'Shaughnessy posted on X announcing an Invest Like the Best episode with Neil Movva, founder of Sail Research. O'Shaughnessy described Movva as having started at Nvidia working on GPUs and now running Sail Research. The episode covers AI inference from software to chips and power, including latency versus throughput, chip pricing, data center strategies, Nvidia lore, and Movva's quoted view that token consumption today differs from dot-com era speculation because users consume tokens immediately. Replies on X largely praised the detailed explanations, while some questioned speculative risks in data center builds and AI funding.
Neil Movva (@neilmovva) started his career at Nvidia, working on GPUs and kernels, and has an unusually deep understanding of inference, from software to chips to power. We spend a lot of time on each of those layers, how they connect, and where the important tradeoffs are.…
Reactions
2 postsPO 
Patrick OShaughnessy@patrick_oshagNeil on why AI spending today is different from the dot-com bubble: "A lot of the investment in networking equipment historically was speculative. We anticipated this future demand for users that never came. What's interesting about token consumption is that it's no longer…
- likes: 110
- replies: 10
- bookmarks: 72
- reposts: 16
IS Ilya Sukhar@ilyasuGreat episode. Favorite part: "I was born in 1997 [...] I remember the time where Cisco was the most valuable company in the world."
- likes: 111
- replies: 3
- bookmarks: 46
- reposts: 8