Qualcomm unveiled its newest flagship smartphone processors at its annual Snapdragon Summit here this week, chips built to run substantial AI models directly on a phone rather than in the cloud — even as a worsening memory-chip shortage threatens to push the price of the devices that use them higher than ever.
The company introduced two new processors, the Snapdragon 8 Elite Gen 6 and a more powerful Snapdragon 8 Elite Extreme Gen 6, both built on a 2-nanometer manufacturing process with a custom Oryon CPU, an Adreno graphics unit and a Hexagon neural processor tuned for on-device AI work. Qualcomm says the top-end Extreme chip can run a 30-billion-parameter mixture-of-experts AI model entirely on the device, without sending data to a server, enabling features such as an always-listening AI notetaker, real-time speaker identification and voice-driven AI agents that respond without a network round-trip.
Smartphone makers including Xiaomi, Vivo, iQOO, ZTE, Oppo, Motorola and Honor are expected to build the new chips into flagship phones arriving as soon as October, with Xiaomi's next flagship line reportedly first in line in China ahead of models from OnePlus, Realme and Asus.
A chip launch shadowed by rising costs
The unveiling comes as the broader smartphone industry grapples with a memory shortage that has driven up prices for the RAM and storage chips that go into every phone. Qualcomm itself is reportedly pushing double-digit price increases on the new chips to its Android partners, a jump industry analysts trace in part to the roughly $30,000-per-wafer cost of the 2-nanometer manufacturing process the chips are built on. Combined with pricier memory, manufacturers pairing the top-tier chip with next-generation RAM and storage could see component costs alone approach several hundred dollars per device.
That cost pressure is expected to show up most visibly in Qualcomm's high-end "Ultra" tier of flagship phones, which analysts expect to carry noticeably higher retail prices than their predecessors when new models ship in the coming months. The dynamic marks a shift from recent years, when component costs were a secondary factor behind flagship phone pricing.
Qualcomm frames the new chips as a competitive answer to rivals also chasing on-device AI, including Apple, whose most recent mobile chip runs a smaller 20-billion-parameter mixture-of-experts model. The company is betting that running larger AI models locally, without a cloud connection, will become a selling point for privacy-conscious buyers and a practical one for users in areas with unreliable connectivity.
Whether consumers embrace the trade-off of higher prices for heavier on-device AI capability will become clearer once phones built on the new chips reach shelves, starting with Xiaomi's launch expected within weeks.
