Qualcomm recently held an analyst event in San Diego, showcasing its advancements in AI technology. The company is focusing on making large AI models compact enough to run efficiently on mobile devices, robots, and cars. Qualcomm’s hybrid AI strategy aims to reduce costs and increase personalization by combining cloud-based and on-device processing. Key techniques include quantization, compression, distillation, and speculative decoding, which allow smaller models to deliver high-quality results. For example, an optimized 8 billion parameter model can perform as well as a 175 billion parameter model. Qualcomm’s AI Hub provides developers access to these innovations, with Snapdragon 8 Gen 3 powered phones running AI tasks up to 40 times more efficiently than traditional data centers. The company plans to launch new Snapdragon SoCs based on Oryon cores this fall and is working on a new version of its AI100 Ultra for data centers. Qualcomm’s automotive business also shows significant growth, with a $30 billion pipeline. Overall, Qualcomm is expanding its AI capabilities across various sectors, positioning itself as a leader in edge AI technology.
Qualcomm’s AI Innovations – Smaller Models, Bigger Impact
Qualcomm is revolutionizing AI by making massive models small enough to run on mobile devices.
1–2 minutes
TOP STORIES
The midterm elections reveal a growing public backlash against data centers, prompting tech companies to rethink their strategies …
Hundreds of instances of the AI hacking tool Artex are exposed online, raising cybersecurity risks …
AI spending growth has stalled as competition drives prices down and firms become more efficient …
Tesla is discontinuing the Cybertruck to focus on producing robotaxis …
Palo Alto Networks is negotiating to acquire Clutch Security for $500 million, focusing on AI access security …
AI models are changing the landscape of cybersecurity, revealing vulnerabilities faster than ever …
