The insight that changes how you think about building with open source AI: the model itself is rarely where the money is made.

As Hugging Face recently detailed, the boom in open source AI models — from Meta’s Llama to Chinese entrants like Qwen and DeepSeek — has shifted the value capture point away from owning the weights and toward everything that surrounds them: fine-tuning pipelines, deployment infrastructure, proprietary data flywheels, and enterprise support.

Where the real value hides

Open source models are becoming mainstream because enterprises, governments, and startups all want lower costs, more control, and more flexibility. But when anyone can download a state-of-the-art model for free, the question becomes: how do companies actually profit?

Hugging Face’s answer is concrete. The value today flows into:

  • Fine-tuning and customization services (adapting a base model to your domain data)
  • Inference infrastructure and hosting (running models reliably at scale)
  • Proprietary datasets and retrieval-augmented generation (RAG) pipelines
  • Consulting, compliance, and security wrappers for regulated industries

In other words, the model is a commodity; the system around it is the moat.

The geopolitics of open weights

A notable point from Hugging Face’s analysis is the rise of Chinese open models. Players like Alibaba’s Qwen and DeepSeek have released competitive base models under permissive licenses. This puts pressure on American open source efforts to differentiate not on raw benchmark scores, but on ecosystem quality, documentation, and enterprise readiness.

For developers, this means the choice of which open model to build on increasingly depends on the support ecosystem, not just the model card. Hugging Face itself is positioning as the neutral platform where any open model can be hosted, evaluated, and deployed — regardless of origin.

What this means for your next project

If you’re building a product on open source AI today, the tactical takeaway is simple: don’t bet your business on model exclusivity. Bet on your data pipeline, your deployment reliability, and your ability to fine-tune faster than your competitors.

The models will keep getting cheaper and more capable. The winners will be those who build the best wrapper around them.

Source: TechCrunch

Source: https://techcrunch.com/video/open-source-ai-is-booming-hugging-face-explains-who-profits/