A comment from the chip industry hints at where consumer AI may be heading. According to a roundup that cites an Alex Heath report, Qualcomm CEO Cristiano Amon says AI companies are starting to make phones, and have asked Qualcomm to support running 100-billion-parameter local language models continuously by 2028. I have only seen this summary, not the full interview.
Why would AI developers want their own phones? A phone is the most personal computer most people own. It holds messages, photos, location and daily habits, and an AI assistant with access to that context could be far more useful than one living in a browser tab. Controlling the device could also give AI companies more control over the user experience and less dependence on other platforms.
The technical challenge is big. A model with 100 billion parameters is large, and running it continuously on a battery-powered device requires efficient chips, lots of memory and careful power management. That explains why chipmakers are central to the story. Qualcomm and its rivals compete to supply the processors that make on-device AI possible.
For consumers, the benefits could include faster responses, better privacy and features that work without a connection. The risks include higher device prices, battery drain and new questions about how much data an always-on assistant collects. A 2028 target also means this is a roadmap rather than a product you can buy today. Watch for announcements from phone makers and AI labs, and for how privacy promises are written.
.png)
.png)
0 Comments