Clem Delangue, the co-founder and CEO of Hugging Face, shared a perspective that stands out amid the ongoing hype around artificial intelligence. Speaking at an Axios event, he argued that while AI as a whole is not in danger of collapsing, the current obsession with large language models (LLMs) might be approaching its peak.
According to Delangue, talk of an AI bubble is everywhere, but the real issue lies in how much attention and investment is being poured specifically into LLMs — the technology behind systems such as ChatGPT, Gemini, and other conversational tools. He believes this segment has been inflated to the point where it could start to deflate as early as next year.
He stressed that LLMs represent only one slice of the much wider AI universe. Fields like biology, chemistry, audio, image processing, and video generation are still at an early stage of innovation, and he expects significant growth in those areas over the next several years.
Delangue also highlighted a misconception fueling the bubble: the belief that a single massive model can serve every purpose for every user. In his view, this one-model-fits-all mindset is unrealistic. He predicts the market will shift toward a diverse ecosystem of smaller, task-specific models designed to solve targeted problems more efficiently.
To illustrate, he gave the example of a banking assistance bot. Such a system doesn’t need the philosophical depth of an advanced general-purpose model. A lightweight, specialized model would be more practical, cheaper to operate, faster, and potentially even deployable on a company’s own infrastructure — something many enterprises prefer for privacy and performance reasons.
Delangue acknowledged that if the LLM market cools down, it could affect Hugging Face to some degree. However, he emphasized that the AI sector is already broad and mature enough that the decline of one segment won’t derail the entire industry, nor will it significantly threaten his company’s long-term position.
He also pointed out that Hugging Face has been conservative with its resources. The company still holds around half of the $400 million it has raised — a sharp contrast to other AI players, particularly those heavily invested in LLMs, who are burning through billions.
With 15 years of experience in the AI world, Delangue says he has seen multiple cycles of hype and slowdown. Hugging Face’s strategy, he explained, is to avoid short-term panic and instead focus on building a sustainable, long-lasting organization that delivers real value.
