Sep 30, 2026 · 54 min
AI spending reshapes markets from chips to software
The $1 Trillion AI Buildout | State of Markets
The episode tests whether AI’s market surge reflects a bubble or the early stages of a far broader industrial and software transformation.
- 1AI demand is driving a capital cycle that spans compute, chips, power, data centers, construction, and manufacturing.
- 2Falling inference costs and agentic workflows could expand AI adoption from experiments into core enterprise and consumer operations.
- 3The opportunity extends beyond foundation models into applications, infrastructure, robotics, biology, health, defense, and autonomy.
Don't miss
The hosts argue that AI’s defining investment cycle reaches beyond software into power, water, roads, factories, and other physical infrastructure.
The brief
David George introduces a State of Markets discussion with Sarah Wang and Santiago Suarez on AI’s expanding economic footprint, from market valuations and enterprise adoption to infrastructure and venture investing.
The central tension is whether AI has created a bubble or a durable earnings cycle; the hosts note that profits have grown even as market multiples have fallen since 2021.
Hyperscaler capital spending could approach $1 trillion annually as usage shifts from occasional prompts to persistent agents, exposing constraints in chips, power, construction, and data centers.
Falling inference costs are making iterative agent workflows more practical, while application companies can capture value by adapting models to proprietary enterprise processes and reliable outcomes.
The discussion closes by widening the opportunity set beyond foundation models to consumer agents, robotics, autonomy, biology, health, defense, semiconductors, and AI-native applications.
Amazon.com, Inc.
Google
OpenAI
AnthropicListen to the full episode and explore every guest, topic, and moment on PodLume.