While artificial intelligence investment continues to soar, a growing chorus of industry analysts are questioning weather the current focus on large language models is sustainable. Hugging Face CEO Clément Delangue is among those suggesting the market might potentially be prioritizing a narrow segment of AI-specifically, the development of generalized “chatbot” solutions-at the expense of other promising applications. This debate comes as venture capital funding for AI startups reached $29.1 billion in the first half of 2024, according to PitchBook data, raising concerns of a potential bubble.
Concerns Rise That Hype Surrounding Large Language Models May Be a Bubble
Recent investment in artificial intelligence (AI) has largely focused on companies like OpenAI and Anthropic, but some industry observers believe the current enthusiasm for large language models (LLMs) may be overblown. Hugging Face CEO Clément Delangue suggests the focus is too narrow, concentrating on a single, albeit prominent, area within the broader AI landscape.
Delangue argues that much of the discussion centers on companies whose core product is LLMs or the data centers required to run them, often offering generalized “chatbot” solutions. He expressed skepticism about these broadly applicable tools. “I believe we are in an LLM bubble, and it could burst as early as next year,” Delangue told Axios. “But LLMs are only a part of the overall AI spectrum. When it comes to AI applications in biology, chemistry, image, audio, or video processing – we are only at the beginning and will see significantly more progress in the coming years.”
Currently, Delangue explained, a disproportionate amount of attention, effort, and capital is being directed toward the idea of creating a single, powerful model capable of solving a wide range of problems for both businesses and individuals. He believes this “one-size-fits-all” approach is unrealistic. Instead, he anticipates a future driven by “many, many models that are more tailored and specialized, and capable of addressing specific tasks.” This shift could reshape the competitive dynamics within the AI sector.
Hugging Face’s Role and Broader Market Trends
This outlook aligns with Hugging Face’s mission to create a platform – similar to GitHub – for hosting and distributing these specialized models. The platform includes large models from industry leaders like OpenAI and Meta (such as GPT models and Llama 3.2), as well as refined versions for specific needs, and smaller models designed for research purposes. This focus is central to Hugging Face’s strategy, providing a logical basis for Delangue’s perspective.
However, Delangue is not alone in identifying this trend. In April, Gartner predicted that “the need for higher accuracy and diverse business workflows is driving a shift toward specialized models tailored to specific functions or domain data.” Regardless of the trajectory of LLM-based projects, investment in other emerging areas of AI is only just beginning.
New Investment Signals Broader AI Potential
Further illustrating this point, former Amazon CEO Jeff Bezos recently announced his co-founding of a new startup focused on applying machine learning to engineering and manufacturing. The venture launched with over $6 billion in funding. While this investment could also prove to be speculative, it clearly demonstrates the wider potential of AI beyond language models.
While some of Delangue’s comments are strategically beneficial to Hugging Face, his message carries broader significance: the term “artificial intelligence” encompasses far more than just LLMs. The technology is still in its early stages, and its full potential remains largely unexplored.
