The Data Exchange with Ben Lorica The Hidden Challenges of Running AI at Scale in Production
9 snips
Mar 12, 2026 Chen Goldberg, EVP of Engineering at CoreWeave and former Google engineering leader, speaks about moving AI from pilot to production. He covers when to choose AI-first clouds, specialized tooling and partner-style support, and why orchestration and hardware assumptions change for large-scale training and inference. Practical challenges like telemetry, resource constraints, and agent workloads are also discussed.
AI Snips
Chapters
Transcript
Episode notes
AI Has Rapidly Shifted From Pilots To Production
- AI adoption moved from experiments to production rapidly across industries beyond Silicon Valley.
- Companies now evaluate assets, data, and partnerships to differentiate rather than only run pilots.
Partner With Specialized AI Infrastructure
- Explore best-of-breed specialized tools when AI workloads need security, reliability, or scale.
- Partner with focused providers to save startup resources and accelerate product development.
Startup Spent Raise On Infrastructure Partnership
- A startup used much of its funding to partner with CoreWeave rather than build infrastructure themselves.
- They preferred outsourcing non-differentiating infrastructure to accelerate product focus.