Sep 5, 2026 · 31 min
Aaron Levie makes the case for open AI models
Aaron Levie on Why Open AI Wins
The debate over open weights now reaches beyond model access to national competitiveness, enterprise economics, and who captures value as AI becomes cheaper.
- 1Open-weight models can broaden innovation and customization while strengthening the wider American AI ecosystem.
- 2Closed labs gain near-term API revenue, but falling token costs could make open models strategically more valuable.
- 3Enterprise AI may depend less on one winning model than on routing tasks, data, and workflows across providers.
Don't miss
Levie argues that model routing will become central as enterprises match changing models to specific tasks, data, and workflows.
The brief
Box CEO Aaron Levie argues that open-weight models accelerate innovation, customization, and competition rather than simply weakening frontier AI companies.
The central strategic question is whether the United States should restrict powerful models or make them widely available before competitors do.
Levie says open models could spread economic value through cheaper inference and a broader American ecosystem, even as frontier labs protect closed APIs.
Box’s evaluations of Opus 5 and Fable shift the discussion from model prestige to cost, reasoning, coding, and agentic tool use.
The conversation ends with a practical forecast: enterprises will increasingly route work across models, data, and workflows instead of choosing one permanent provider.
Sofia Puccini
Box
Jen-Hsun Huang
United States
GPT-6 AstraListen to the full episode and explore every guest, topic, and moment on PodLume.