Clem Delangue, CEO of Hugging Face, told TechCrunch’s Equity podcast that open-source artificial intelligence is booming. Over recent years Hugging Face has evolved into a GitHub-like hub for AI, where builders can share and download models and datasets. The company’s offerings are now used by roughly half of the Fortune 500.
From frontier APIs to open-source models
Delangue described a recurring pattern: companies often start by using frontier APIs, but as they scale the cost pressures push them toward open-source models. According to him, economics and the demands of scaling explain much of the migration to open approaches.
Concern about concentration of control
In the interview Delangue expressed concern that a small number of large companies could end up controlling large parts of the AI ecosystem. The open vs. closed-source debate was discussed in the context of Anthropic’s halted Fable release, which Delangue sees as part of a broader conversation about future competition and access.
Chinese labs’ role in open-model supply
Delangue noted that Chinese labs produce the majority of the open models downloaded in the United States. He said this is an issue worth addressing, but it should not be taken as a reason to distrust open source itself.
Capital-efficiency over big fundraising rounds
Hugging Face favors capital efficiency rather than following the typical Silicon Valley fundraising playbook. Delangue mentioned that the company turned down a large investment offer from Nvidia last year, citing the importance of maintaining capital efficiency and independence.
Robotics as an urgent case for openness
Delangue argued that robotics may be an even more urgent domain for transparent, open AI than chatbots or coding tools, because robots see much of home and family life, making openness and control especially critical in that context.
Where to hear the full episode
The full conversation is available on TechCrunch’s Equity podcast, which can be accessed on YouTube, Apple Podcasts, Overcast, Spotify and other podcast platforms. Equity also maintains presence on X and Threads at @EquityPod.
