Qualcomm's Next Hexagon NPU: 50% More Shared Memory, 30B MoE Models on a Phone
ryanshrout · x · 2026-09-11
Third pre-Summit deep dive: Qualcomm's next Hexagon NPU adds an Element Accelerator combining vector and scalar extensions for transformer workloads, a shared memory pool 50% larger than Snapdragon 8 Elite Gen 5, and support for MoE models up to 30B parameters.
Claimed gains vs 8 Elite Gen 5: up to 50% higher INT4 pre-fill performance, time-to-first-token as low as 1.5s.
The article argues MoE on a phone is fundamentally a storage problem: a 30B MoE keeps tens of billions of parameters available while activating only 3B routed parameters per token, keeping compute within a phone power budget while behaving like a much larger model.
Related event: Qualcomm's Next-Gen Hexagon NPU to Run 30B MoE Models On-Device(3 posts)→
More from Embodied
- HandsOn benchmark: 66 verified tasks and 18,000 human demos for dexterous manipulation — stingning · 2026-10-27
- DexPort: a handheld interface with a real robot hand for dexterous-hand data collection — pabbeel · 2026-10-27
- ROS names Alejandro Hernández Cordero new project lead, eyes 2027 release — ___Mufasaa · 2026-10-27
- ROS Meetup coming to Google's NYC office Oct 29 with speakers from Tangent, Cohesive Robotics, Intrinsic — kscottz · 2026-10-27
- Tesla officially expands robotaxi service area across much of Houston, Texas — Polymarket · 2026-10-27
- Walden Robotics teaches robots to call for help, blending remote assist with autonomy — adnothing · 2026-10-27