Hugging Face CEO Identifies New Trend: Enterprises Shift from 'Renting' to Building Their Own AI
Clem Delangue, CEO of open-source AI giant Hugging Face, reveals companies are transitioning from API-based 'rented' AI to building and customizing their own open-source models for superior flexibility and efficiency.
The trend of open-source AI is growing exponentially, and companies are starting to look for new alternatives beyond 'renting' AI models from major providers. This is the latest perspective from Clément Delangue, co-founder and CEO of Hugging Face, a platform akin to the 'GitHub of AI,' currently used by roughly half of Fortune 500 companies. Delangue told TechCrunch that he observes a recurring pattern: initially, companies often start by using closed-source AI models via APIs (Application Programming Interface), which he likens to 'renting' an entire car. It's easy and quick, but after a while, organizations begin to realize the limitations, including rising costs with increased usage, a lack of flexibility for customization, and concerns about sending data for external processing. This turning point leads many companies to embrace a more 'ownership' approach by developing further with open-source AI models. Delangue likens open-source models to 'engines' that companies can use to build their own 'cars' or highly specific applications directly. Organizations can choose smaller, specialized models, which results in faster performance, lower costs, and crucially, the ability to run them on their own infrastructure, providing better control over data and security. Hugging Face plays a crucial role in driving this trend, serving as a central hub for developers and organizations to access, share, and collaborate on hundreds of thousands of AI models, datasets, and tools. The platform has lowered barriers to high-quality AI technology, enabling companies to build their own AI capabilities instead of relying on services from just a few major players. Delangue believes this is the true future of enterprise AI.
This trend is highly significant for Thai businesses and developers, as it opens the door to building AI that is more easily and cost-effectively adapted to the Thai context, without sole reliance on foreign services.