Qualcomm is assembling chips, connectivity and acquisitions into a full-stack platform for industrial AI, according to RCR Wireless News.
The move comes as it competes with Nvidia to decide who actually profits from putting artificial intelligence to work in physical factories and machines.
The newsletter's Telco Diary column draws on reporting from Qualcomm's Dragonwing IoT Day in San Diego.
It argues that the two companies differ in emphasis rather than being opposed.
Nvidia wants to own the centralised computing power that drives its AI "factories".
Qualcomm wants to control the edge, meaning the computing, connectivity and systems inside devices.
That would allow intelligence trained in the cloud to act in real time inside robots, drones and industrial sensors.
Qualcomm's Dragonwing portfolio has moved through IQ6, IQ8 and IQ9 processors launched in 2024 to an IQ10 series in 2026 aimed at advanced robotics.
And the company has set a fiscal-2029 target of more than $14 billion in IoT revenue, including roughly $8 billion from industrial, networking and robotics.
Its newer IQ-2390 chip, built for industrial conditions from –30°C to +115°C with availability committed through 2036, illustrates the long product cycles and deterministic reliability industrial buyers demand, a harder problem than shipping raw AI capability.
That tension sits behind Nokia and Ericsson's retreat from private 5G as a standalone pitch; Ericsson has folded its enterprise-wireless unit into its core networks business, saying the move streamlines development rather than abandons the technology
Read more about Semiconductors
