The New Stack Podcast Google AI Infrastructure PM On New TPUs, Liquid Cooling and More
21 snips
May 13, 2025 Chelsie Czop, Senior Product Manager for AI Infrastructure at Google Cloud, dives into cutting-edge developments in AI hardware. She discusses the impressive new Ironwood TPUs, boasting 42.5 exaflops, and the advancements in liquid cooling, essential for managing heat. Chelsie clarifies the ongoing debate between using TPUs or GPUs, noting significant performance boosts for some users. Moreover, she highlights the collaboration with DeepMind to stay ahead of evolving model architectures and the sustainable innovations shaping the future of data centers.
AI Snips
Chapters
Transcript
Episode notes
Synergy of Hardware and Software
- Close collaboration between hardware and software teams is vital for TPU usability.
- Software like Google's Pathways framework unlocks hardware potential for large model training and inference.
Bridging Hardware and Model Evolution
- Models evolve rapidly while TPU hardware releases follow a steady annual cadence.
- Software optimizations help bridge the gap between fast model changes and slower hardware iterations.
Faster Hardware Cadence at Google
- Chelsie Czop came from IBM, where hardware design took 3 years, to Google, where TPU development cycles are much faster.
- She admired Google's philosophy of designing hardware to fail and relying on software to maintain reliability.