Cofounder and CEO, Huggingface
Hugging Face CEO Weighs In on Anthropic AI Model's 'Dangerous' Label
As government scrutiny over Anthropic's Mythos AI model takes center stage in the public debate over AI's national security concerns, Hugging Face CEO Clem Delangue says being labeled "too dangerous" may actually be good marketing for frontier AI firms. He joins Ed Ludlow on "Bloomberg Tech" to explain.
--------
Like this video? Subscribe to Bloomberg Technology on YouTube:
/ @bloombergtechnology
Watch the latest full episodes of "Bloomberg Technology" with Caroline Hyde and Ed Ludlow here:
• 'Bloomberg Tech' With Ed Ludlow
Get the latest in tech from Silicon Valley and a...
About Clement Delangue
Clement Delangue, cofounder and CEO of Hugging Face, made a series of media appearances in mid-2026 discussing open-source AI, regulation, and robotics. On the Equity Podcast, Delangue said he has repeatedly observed companies starting with proprietary frontier APIs and later switching to open-source models for cost reasons. He described the potential risk of a few companies controlling AI, stating that keeping models behind closed doors creates an "asymmetry of power" and argued that the world can be made safer by "leveling up the playing field" through open-source competition. Commenting on government scrutiny of Anthropic's Mythos model, Delangue said that being labeled "too dangerous" can serve as effective marketing for frontier AI firms and noted that it is "fair for the US government to at least try to get more transparency." He cautioned against spreading regulatory constraints to smaller entities such as startups and academia, which lack the resources of large AI companies. Delangue also promoted open-source approaches in robotics, citing Hugging Face's release of a small robot called "Rich Cimini" and stating that transparency and control over robots that interact with people require open-source development.
Transcript (10 segments)
H
Host0:00
Clem, I want to start with that post. You put on the idea that that designation is good marketing. Could you explain a little bit more why you think that?
C
Clement Delangue0:12
Well, first I want to point out that there are people saying that it's not completely unfair for them to start being regulated, given that they've been doing what we call doom marketing for quite many years now. If you remember GPT-2, that was really, I think 5 or 6 years ago was already deemed too dangerous to release. So I feel like it's fair for the US government to at least try to get more transparency about what these models can and can't do. And I think it's important to remember that these are gigantic companies, fastest growing in the world on their way to maybe be the most valuable companies by the end of this year. So I feel like they can take these kinds of regulation and interaction with the US government. I think one thing that we need to be careful of is to spread these constraints and regulations to the rest of the ecosystem like startups, little tech, university, academia who certainly don't have the same kind of legal and policy capabilities that these companies have.
H
Host1:22
I want to get to that distinction of regulating sort of the Frontier Labs and APIs and open source in just a minute, but to end the conversation on that post, do you think that the US government have handled this correctly with Anthropic and Mythos and Fable?
C
Clement Delangue1:40
Well, I think it's important to get more transparency for everyone, right? The government needs to have more transparency about what these models are capable or not capable of doing, which is hard to do through APIs. Right? Because there are guardrails, there are limitations. That makes it hard to actually know what they're capable and not capable of doing. So I think it's quite fair for the US government to try to get more transparency on these very complex black box system. Obviously, it's hard to do that perfectly, given how fast the field is moving and how uncertain some of these risks are. But I think it's not unfair overall.
H
Host2:24
You argue that there should be a distinction how governments regulate frontier AI and open source. Why? Why should open source be treated differently? What is it that the ecosystem needs differently on the open source side?
C
Clement Delangue2:41
Well, first, in terms of capabilities, I think it's broadly accepted that the most dangerous capabilities are concentrated in these frontier AI labs where the majority of open source models are more spatially smaller, more broadly beneficial models that aren't really creating more risks. So there's a difference in terms of capabilities. There's a difference in terms of transparency. You know, open source models are really easy to evaluate and to look at what they can and can't do from the get go because you get the model, so you can test it really easily versus an API, which is really hard to do. So you need more kind of access to really assess them. And also the provenance is different. While open source is like a nickel system of smaller companies, smaller organizations versus the fancy labs which are a couple of companies concentrating a lot of power behind closed doors.
H
Host3:52
There's been a lot of press coverage of late and other data points like your own revenue run rate. Right. Open source is your lifeblood, but the open source models seem to be competitive in a market against closed models. What else would you point to to evidence that?
C
Clement Delangue4:10
Well, it's not just us, right? The whole ecosystem right now is booming. We see tons of companies doing super well, like the inference providers Neo Cloud or Boosting. Fantastic growth. A million new models and data sets have been shared on Hugging Face just this past quarter. And you see that in usage too, the usage of open source by American companies, American startups, American small companies is also booming. So this is really exciting because it basically enables and empowers more people to really own AI themselves rather than renting it with APIs. So this is quite exciting.
H
Host5:01
Another area of real focus of the show of late has been robotics. We had D.P. Tyler from Nvidia on last week talking about HALOs, Hugging Face, also looking a lot at robotics. What I would be really grateful for is to understand the role that open source is playing in robotics and physical AI, but also the benefit of leaning on open source in that field.
C
Clement Delangue5:23
I feel like it's even more important than in general. Just because if you think of having a robot at home that is interacting with your environment with your kids. I just got two daughters a few months ago. So when I think about robots interacting with my daughters, I don't want to have to trust a black box controlled just by one megacorp being able to do anything. I want to have some transparency about what's going on on this robot. How is it built? How does it decide to interact one way or another? And for that, open source is the only way, right? It gives you transparency, it gives you control. It creates an ecosystem of different companies that are going to be able to build the robots. So that's why we're excited about it. We created this little robot called Rich Cimini, and we've been really surprised by how into it people have been. We shipped over 10,000 of them all over the world in the past few months. And that's just testament to how excited people are by more open, more transparent, more open source approaches to robotics.