Qualcomm and Hugging Face have expanded their partnership to run open AI models smoothly from device to cloud.
Qualcomm Technologies and Hugging Face have expanded their strategic partnership to bridge edge devices and cloud networks. The collaboration allows Hugging Face’s 16 million developers to optimize and deploy open AI models seamlessly across Qualcomm’s entire hardware portfolio—spanning Snapdragon mobile processors, Dragonwing on-prem appliances, and the newly launched Dragonfly data centre racks. The Qualcomm users get access to Hugging Face PRO for premium storage, compute, and collaboration.
To deliver balanced latency, cost, and performance, the initiative focuses on three key pillars. Firstly, Data Centre Integration maps Hugging Face’s global storage and inference pipelines directly onto energy-efficient Qualcomm Dragonfly hardware. Secondly, Zero-Manual Deployment utilises an automated Hugging Face Agent alongside Modular software tools to instantly set up and optimise over 3 million open models with zero manual integration work. Agentic AI Orchestration supports a distributed, hybrid framework where intelligent agents can fluidly move workflows between local device chips and data centre servers based on local privacy and compute constraints, being the third pillar.
Qualcomm CEO Cristiano Amon emphasised that the merger of high-performance, low-power hardware with a massive community effectively democratises advanced AI. Hugging Face CEO Clément Delangue added that as the tech sector shifts towards private, affordable, and localised open models, this partnership enables global developers to run applications smoothly across the complete compute continuum. The data center pillar originally highlighted creating a “direct path from model experimentation to production deployment.”
