# Qualcomm CEO Says Phones May Need 100B-Parameter AI Models by 2028

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/29601
Published: 2026-10-10T16:57:13.000Z
Updated: 2026-10-10T16:57:13.000Z
Section: Technology

> Cristiano Amon said AI companies want smartphones to run models continuously at that scale, while Qualcomm’s newest platform supports more than 30 billion-parameter models.

Qualcomm CEO Cristiano Amon said AI companies want smartphones to run 100-billion-parameter models continuously by 2028, a target that would raise demands on mobile chips, memory and battery life.

Amon said in a Sept. 23 interview that companies were asking for “100 billion model parameters running on your device.” He added that some consumers want AI and their data to remain on the device and under their control.

He gave the 2028 timing in a later interview on Oct. 8, describing the target as a request from AI companies rather than a Qualcomm product launch date or delivery commitment.

Parameter counts offer a rough indication of how large an AI model is. As that number grows, phones generally need more memory, processing power and energy, complicating continuous use.

Qualcomm’s Snapdragon 8 Elite Extreme Gen 6 mobile platform is designed to run more than 30 billion-parameter mixture-of-experts models directly on a device. The 100-billion-parameter target is therefore more than three times the model scale Qualcomm currently advertises for its newest mobile platform.

Qualcomm has not announced a smartphone chip capable of running 100-billion-parameter models continuously.

Qualcomm is promoting the Snapdragon 8 Elite Extreme Gen 6 and Snapdragon 8 Elite Gen 6 platforms for on-device agentic AI. The company defines agentic AI as software that understands intent, anticipates needs and takes actions for users.

“We now have to design chips to allow the phone to be working, even when you're not operating it, and with an all day battery life and a lot of AI at the edge,” Amon said.

Qualcomm said the two platforms are made using a 2-nanometer process. Their emphasis on local AI processing reflects efforts to keep models and personal data on smartphones while supporting background functions and user-directed actions.