Tech Brew Ride Home – (BNS) Hugging Face Founder Clément Delangue

Episode Date: October 23, 2025
Host: Techbrew (Morning Brew)
Guest: Clément Delangue (Co-founder and CEO, Hugging Face)

Episode Overview

In this episode, Clément Delangue, founder of Hugging Face, shares the story of his entrepreneurial journey—from his early tech days in France to building one of the most influential open-source AI platforms. He discusses the startup and tech scenes in France, New York, and the US; unpacks the pivotal moments in Hugging Face’s development; and offers insights into the state and future of open-source AI globally.

Key Discussion Points & Insights

1. The Importance of Open Models and Decentralized Power

[00:26] – [01:36]

• Clément addresses the current disparity between open-source AI models in China vs. the US and the importance of democratizing model development globally, warning against the concentration of power.

  “I think we do care for reasons of concentration of power. I think ultimately we want actually not just China and the U.S. but any country to be able to produce their own AI models… to make sure that power isn't concentrated to avoid some of the biases contained in different models from different countries.”
  — Clément Delangue [01:08]

2. Origins: Early Computing and eBay Journey

[02:02] – [07:52]

• Clément reminisces about his first computer in 1998, sharing how sibling rivalry sparked early online ventures:

  “…fighting with my siblings… because you couldn't, couldn't receive a phone call when you were, when you were on the Internet, obviously. Remember the noise of the router … we started … Internet trading, like buying on one platform, selling on another…”
  — Clément Delangue [02:34]

• Stint at eBay during business school: landed an internship after being a big seller; served as a seller advocate helping shape the mobile platform.

  “I was kind of like a little bit the seller representative in a way, like the chief of the union of the sellers, like trying to move the product into something that can make sense for more sellers.”
  — Clément Delangue [06:35]

3. First Startups and Lessons in Community

[08:19] – [10:48]

• Founded Unishared (collaborative/open student note-sharing platform): learned about the challenges of visions vs. user behavior.

  “…it wasn't always the same people wanted to do both [collaborate and share openly]…I thought that overnight I could change kind of like everyone's behavior versus now... I've learned that you have to be more progressive...”
  — Clément Delangue [10:27]

• Lesson: Balance idealism with practical adoption and user signals.

4. First Steps into AI: Moodstocks and Mention

[11:04] – [14:17]

• Joined Moodstocks (early computer vision startup):

  “He was calling it computer vision at the time, but it's basically AI now... you just point your phone at the product and with AI, you can recognize the product. It's like you crazy. What, what is that?”
  — Clément Delangue [12:01]

• Learned deep tech from zero:

  “For the first months, I basically understand nothing about what they're talking about.”
  — Clément Delangue [13:09]

• At Mention (social listening/alerts): First experience expanding into the US market.

5. Moving to NYC & Meeting Hugging Face Co-founders

[14:17] – [17:31]

• Chose NYC for its international vibe and startup energy:

  “…it felt to me like the perfect merge of some of the things that I liked from Europe and some of the things that I liked from the US... I didn't only fell in love with the tech here, but also with the city really quickly.”
  — Clément Delangue [15:25]

• Already knew Hugging Face co-founders Julien & Thomas from Paris startup circles. Met Betaworks (early investor) in NYC.

6. The Origins and Pivot of Hugging Face

[17:31] – [28:43]

• Started as an “AI Tamagotchi”/chatbot — more playful & interactive than Siri or Alexa.

  “...this is very productivity driven, you know, like, this is designed to just give you the weather... maybe there's something more interactive, something more fun.”
  — Clément Delangue [19:05]

Naming the Company:

• Hugging Face was a “joke” about IPO’ing with an emoji ticker, but fans embraced the emoji:

  “...the hugging face emoji, people start to put it everywhere…do T shirts, do the sign...”
  — Clément Delangue [20:44]

The Transformative Pivot

• Internal tooling to keep up with AI research becomes the real product.
• Thomas Wolf ports Google’s new BERT model to PyTorch “over a weekend,” shares it on Twitter, and it goes viral.

  “…Julian and I was like, okay, you know, have fun. You do you if you think it's exciting. I was barely understanding, to be honest, what it was…”
  — Clément Delangue [25:10]

• Community starts using platform to share and distribute models; this compels the pivot.

On Open Collaboration:

• Hugging Face reflects Delangue’s collaborative/open vision, like Unishared did, but for AI developers.

  “…we found the same values in Hugging Face with a different approach and different targets.”
  — Clément Delangue [28:30]

7. Building in Tune With Community Signals

[29:03] – [31:26]

• Start from conviction & intuition, build and release, refine based on community adoption.
• Willingness to experiment with many products/libraries, knowing most won’t take off but some might.

  “Is because we have intuitions, we have ideas and then we release it for the community...some of them are going to prove useful and are going to get adoption.”
  — Clément Delangue [31:07]

8. Navigating the Big Pivot: Investor & Team Relations

[31:31] – [33:28]

• Importance of picking open-minded, generalist investors who support pivots.
• Bring investors onboard after evidence of traction, not during early experimentation.

  “...wait to kind of like bring your investors on board when you have quite high validation and conviction that you're going to pivot...”
  — Clément Delangue [33:08]

9. The AI “Switzerland” Approach and Hugging Face’s Business Model

[33:43] – [38:31]

• Deliberately structured cap table with many cloud/chip investors (Google, AWS, AMD, Nvidia, etc.) to maintain neutrality.

  “So instead of having one player, one investor being the biggest and representing the majority... we decided to include a lot of them.”
  — Clément Delangue [35:28]

• Business model: freemium, with open source/public core and paid enterprise features (Enterprise Hub, Private Hub, custom endpoints).
• Over 50,000 paid subscriptions; big corporate customers pay for advanced management/features.

10. How Should AI Be Evaluated?

[38:31] – [43:03]

• No single evaluation; overreliance on leaderboards is misleading.

• Advocates a threefold approach: social validation (community likes/activity), public benchmarks (task-specific), and personal benchmarks on your own data/use case.

  “...combine these three factors, social signaling, public evaluation, and private evaluation... that's when you can take very informed decisions about what to use...”
  — Clément Delangue [41:02]

• Envisions a future where there are as many AI models as code repositories because specialization and customization will matter more than generalist APIs:

  “...our vision is that there’s going to be very much a diversity of models, the same way you have diversity of code repositories.”
  — Clément Delangue [42:16]

• Hugging Face sees a million new repositories in last 90 days—one every nine seconds.

11. State of Open Source AI: Geopolitics, Policy, and Efficiency

[43:41] – [50:57]

• The US lags China in open source model release.
• Advocates for US policymakers: fund open datasets, distribute compute resources more equally, celebrate openness rather than fear it.

  “...I would send a signal in the US and celebrate when companies are sharing openly... we have to change a little bit our mindsets in the US and try to celebrate when companies, organizations are sharing their research...”
  — Clément Delangue [47:46]

• Open source is more energy efficient—labs don’t redundantly spend compute on the same models, as “closed” labs do.
• Predicts open source will win for cheaper, faster, more specialized models; expects “a big model for ChatGPT,” but “everything else…smaller, faster models based on open source.”
  — Clément Delangue [51:04]

12. The AI “Bubble” and Macro Tech Trends

[51:10] – [52:33]

• AI usage among companies is solid, not “bubbly.”
• If there’s a bubble, it’s in LLM compute—massive spend to train general models, not in AI as a whole.

  “From our standpoint, the foundation in terms of usage from companies are very strong… and I think it's not going to go anywhere.”
  — Clément Delangue [51:52]

13. Reflections on the French/European Startup Ecosystem

[53:05] – [54:23]

• Startup opportunity is growing, with Europe/French tech beginning to raise bigger rounds and scale faster (citing Mistral).
• Suggests Europe should focus on its scientific/industrial strengths (climate, fashion, luxury, etc.), not just copy the US/China.

14. Why New York is a Great Place for AI Startups

[54:23] – [55:51]

• NYC’s diversity (people and industries) is a tech superpower, especially for “AI for x” businesses (finance, fashion, chemistry, etc.).

  “It feels like a truly international city... gives you kind of like a diversity of opinion and support in terms of domains. If I was thinking of starting a company... all these are the domains which, in my opinion, are the most exciting domains of AI.”
  — Clément Delangue [55:00]

Memorable Quotes & Moments

• On Open Source AI’s Importance:
  “In my opinion that's one of the most important geopolitical topic of the decades. Personally I'm not in politics, but if I was a politician I would make it a priority to make sure that the US has strong open source foundations for everyone to use.” — Clément Delangue [45:03]

• On Naming Hugging Face:
  “...the first joke was, what if we're like the first company to go public with an emoji instead of the three letter tickers...” — Clément Delangue [20:15]

• On Platform Pivot:
  “At the time, I think we get maybe a thousand likes on Twitter, which for us at the time we were like, okay, we broke the Internet. What's happening? Because of course we were kind of like random French guys not having much of visibility or recognition online.” — Clément Delangue [25:29]

• On User Behavior:
  “I learned that sometimes when you have values when you have a vision, it's important to keep it, of course, and to build a platform that makes it happen. But maybe not to go too extreme too fast, and rather help the users progressively get to the point that you'd like them to be.” — Clément Delangue [10:25]

Notable Timestamps

• Intro/open vs closed models: [00:26]-[01:36]
• First computer/eBay work: [02:02]-[07:52]
• Unishared/startup lessons: [08:19]-[10:48]
• Moodstocks/Mention/first AI exposure: [11:04]-[14:29]
• Move to NY/founding team: [15:00]-[17:31]
• Hugging Face origin/name: [17:31]-[20:53]
• Pivot to open tools/community: [23:37]-[28:39]
• Community-driven building: [29:03]-[31:26]
• Investor pivot strategy: [31:31]-[33:43]
• Neutral ‘Switzerland’ strategy/business model: [33:43]-[38:31]
• AI evaluation philosophy: [38:56]-[43:03]
• Geopolitics of open AI: [43:41]-[50:57]
• AI market “bubble” chat: [51:10]-[52:33]
• European ecosystem/French tech: [53:05]-[54:23]
• Why NYC for AI: [54:23]-[55:51]

Conclusion

This episode provides a deep dive into the values and strategic pivots that shaped Hugging Face into a central force for open-source AI, as well as Clément Delangue’s nuanced views on tech culture, international competition, and the importance of community-led product development. The conversation is packed with startup wisdom, global AI policy reflections, and practical advice for founders and policymakers alike.

For listeners (and non-listeners), this summary captures the substance, stories, and spirit of the conversation, highlighting the key insights and trajectories shaping Hugging Face and open AI today.