Interview
How NVIDIA makes AI infrastructure more productive
Shruti Koparkar
Senior Manager of Product & Go-to-Market, NVIDIA
42 minHosted by Kim IsenbergRecorded before GTC Berlin
Watch on YouTubeAbout this conversation
What makes an AI factory a good investment? NVIDIA’s Shruti Koparkar joins Kim Isenberg to discuss GPU productivity, the role of software in extending the useful life of infrastructure, and the trade-offs operators face when deploying AI systems.
They break down CUDA and CUDA-X, explain how inference frameworks fit into the stack, and discuss performance per watt and how the ecosystem optimizes new models for deployment.
Topics
- What makes an AI factory a good investment
- GPU productivity
- How software extends the useful life of infrastructure
- The trade-offs of deploying AI systems
- CUDA and CUDA-X
- How inference frameworks fit into the stack
- Performance per watt
- How the ecosystem optimizes new models for deployment
Kim Isenberg is co-founder and Editor-in-Chief of Superintelligence, an AI newsletter published every weekday.
Read the NewsletterMore interviews
- Jacky GuHow XPeng builds robots and EVs
- Dr. Heiko KlarlWho controls your AI agents?
- Christian MonbergWhy your AI pilot never ships