264001
eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJid2V2ZW50cy1wdWJsaWMtYXBpIiwiYXVkIjpbInB1YmxpYy13ZWIiXSwic3ViIjoiaHR0cHM6Ly93d3cuZ29vZ2xlY2xvdWRldmVudHMuY29tIiwiaWF0IjoxNzc1NDE1MzQ5LCJuYmYiOjE3NzU0MTUzNDksImV4cCI6MTc3NTQxODk0OSwiY29udGV4dCI6eyJwYXlsb2FkIjp7ImlkIjoiIiwicmVnVHlwZU5hbWUiOiIiLCJldmVudF9pZCI6IjI2NDAwMSIsImNfMzI5MDQxNiI6IiJ9fX0.bLxTBm1uJxl0oFwTgHjjLto_A6j3fj7ieGHzsfVRn8c
phase 5 session info modal
Day 1 – April 22, 2026
Breakouts
BRK2-172 • Architecture, Compute • Technical
Scaling Claude: Inside Anthropic's TPU strategy and architecture
location_on Mandalay Bay I
schedule 11:00 AM - 11:45 AM
The “compute race” isn’t just about chip count – it’s about delivering the most intelligence per dollar. As training costs climb, efficiency at the silicon level has become the ultimate competitive advantage. Join us to learn why Anthropic, one of the world’s leading AI labs, is betting on a massive scale-out of Google’s Cloud Tensor Processing Units (TPUs). We’ll go beyond the hype to discuss the real-world engineering and economic decisions behind training the next generation of Claude. Discover how to leverage TPU performance-per-watt to turn infrastructure into your competitive edge. Read more
This agenda widget contains logic for handling same sessions available at multiple times
GTM Tracking Header name present
Session
contains logic for gtm tracking when modal open