AI API layer faces rapid commoditization
The guest argued that the pure API layer for AI models will commoditize rapidly because switching costs are zero and enterprises can easily hot-swap models based on automated evaluations.
The argument
As enterprises build robust internal evaluation systems (evals) for specific workflows, they can easily benchmark and distill frontier models into cheaper open-source or custom fine-tuned models, destroying the pricing power of standard APIs.
The thesis, stress-tested
✓ What validates it
- ✓Rapid decline in API pricing per token across major model providers
- ✓Increased enterprise adoption of open-source models like LLaMA for core production workflows
▸ Risks discussed
- ▸Frontier models maintaining such a massive capability gap that distillation becomes unviable
- ▸High complexity in managing custom open-source deployments for smaller enterprises
Hear it yourself
"I could definitely see one of them being a $10,000,000,000,000 company, maybe even significantly higher. How much does it cost to hire a high quality AI researcher? Oftentimes, it would be in the tens of millions of stock per year. This is 20 VC with me, Harry Stebbings. Now joining me in the hot seat today, we have Brandon Foudy, co founder and co CEO of Macaw, one of the fastest growing AI companies valued at over $10,000,000,000 today, doing over $1,000,000,000 in revenue. Now, Brandon's done quite a few shows before and so my question was, how do I get answers that he's never given"
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE