Qualcomm said its phone chip can run 30 billion parameter models locally.
Its new top-end phone chip can run AI models of more than 30 billion parameters on the handset, Qualcomm announced.
Parameters are the internal values a model learns, and their number is a rough measure of model size.
At its Snapdragon Summit in Maui, the company unveiled the Snapdragon 8 Elite Gen 6 and the Snapdragon 8 Elite Extreme Gen 6.
The Extreme version has 50 percent more shared memory. Qualcomm says that lets a model of this size run on the device instead of in a data centre. The AI part of the chip is 35 percent faster than last year's and delivers 33 percent more work per unit of power. Those are Qualcomm's figures, with no published test conditions.
Motorola and Xiaomi are among the first manufacturers to use the chips. ZTE is too.
The claim shows what the chip can do. Qualcomm has not established that phones are already shipping.
The governance question arrives before the phones do. Running the model locally removes the remote call, though the app around it can still send usage data, including prompts and answers. Unless the app keeps a record, your central team may have no log of it.
When a chip like this reaches company phones, whose policy covers what the model reads?
Image: AI-generated.
Sources
Our file on Qualcomm
All 2 Qualcomm storiesFor your next steering committee
- 1Ask which AI tools already touch your ERP or finance data, including the ones nobody approved.
- 2Ask what each tool is allowed to change, and who reviews what it changed.
- 3Ask how the result is measured, and who decides to switch a tool off.
Issue 01 of the ERP LEADERS brief puts a timeline, the three-number review and six questions on one page. Read issue 01
Welcome back. · Issue 01
