AI + a16z AI’s Capital Flywheel: Models, Money, and the Future of Power
138 snips
Feb 24, 2026 
Sarah Wang, a16z general partner focused on AI/model investing, and Martin Casado, a16z general partner expert in infrastructure and networks, discuss the unique capital dynamics reshaping AI. They cover why fundraising now directly drives capability, the blurring of apps and infrastructure, the risk of model firms out-raising their ecosystem, talent wars, custom hardware economics, and two divergent industry futures.
AI Snips
Chapters
Transcript
Episode notes
Two Divergent Futures For AI Markets
- Two possible industry futures: fragmentation with many specialized apps, or concentration where a few SOTA models dominate by scale.
- Casado frames the choice as either infinite market growth or oligopoly driven by capital and generality.
Character Used Product Revenue To Fuel Research
- Sarah describes Character's path: Noam left Google to ship products and used product-driven data collection to pursue AGI.
- She explains the trade-off founders face between product revenue and allocating GPUs for long-term research.
Custom ASICs Become Economical For Huge Training Runs
- At scale a $1B training run can justify a custom ASIC because inference savings compound enormously.
- Casado notes saving ~20–50% on compute can economically validate chip tapeouts for large runs.