Baseten argues that raw GPU compute is a commodity, but inference with a software layer on top creates strong customer retention. The company reports 400% annual net dollar retention and zero churn among its top 30 customers. The core thesis is that the software layer for inference is the real differentiator, and that access to inference compute is becoming a strategic advantage — a race even AI labs are pursuing by locking in compute resources.

1m watch time
15 Impressions