sameer dholakia

Frontier closed models will continue extending their lead through recursive learning improvements, while open-source ecosystems flourish in parallel; IP protection for closed models is needed but competition from Chinese open-weight models is net positive.

Fireworks AI is scaling at unprecedented speed ($100M ARR in <1.5 years) led by a deep technologist CEO from Meta's ML framework team and a proven go-to-market president (ex-Salesforce COO), positioning it to win the open-model inference serving market.

Frontier model providers like Anthropic will continue extending their lead through recursive learning improvements, but need better IP protections against distillation by Chinese competitors to maintain moat.

Fireworks AI is achieving unprecedented growth ($100M ARR per month) by combining deep technical leadership from Meta's ML framework team with world-class go-to-market execution from former Salesforce COO George, positioning it to win the open-model inference market.