Qualcomm Unveils Snapdragon 8 Elite Gen 6 Chips with Support for 30B-Parameter AI Models

Qualcomm introduced two flagship smartphone processors, the Snapdragon 8 Elite Gen 6 and the Snapdragon 8 Elite Extreme Gen 6, focusing on localized artificial intelligence workloads and pro-level media processing at its annual Snapdragon Summit, according to a report by TechCrunch (https://techcrunch.com/2026/09/22/qualcomm-launches-two-new-smartphone-chips-with-emphasis-on-ai/).

Both chipsets incorporate upgraded sensing hubs capable of executing small models containing up to 200 million parameters directly on mobile hardware. Qualcomm stated that this capability allows devices to operate an end-to-end voice-in, voice-out AI agent locally, transcribe audio while distinguishing between speakers, and construct memory from user interactions to automate routine tasks.

To handle larger workloads, the standard Snapdragon 8 Elite Gen 6 includes a revised neural accelerator designed for energy-efficient execution. The higher-tier Extreme edition is engineered to locally run mixture-of-experts (MoE) models of up to 30 billion parameters, activating only task-relevant parameter subsets during inference. By comparison, Apple announced a 20-billion-parameter MoE model for its third-generation foundation model platform at its Worldwide Developers Conference in June.

On the imaging front, the new processors provide pixel-level camera controls, motion understanding, and improved stabilization. The Extreme model supports high-resolution capture, including 8K recording at 60 frames per second, 4K slow-motion at 240 frames per second, and the Advanced Professional Video (APV) recording codec. Both processors also feature machine-learning-driven noise reduction and vocal enhancement, alongside Qualcomm's voice bubble technology to isolate speech during calls.

At the summit, Motorola announced that its upcoming Motorola Signature 27 smartphone will run the Snapdragon 8 Elite Extreme Gen 6, with general availability scheduled for later this year. The product launches reflect a continuing shift across the electronics sector toward concentrating AI functionality inside flagship smartphones rather than specialized standalone hardware.