-
QCOM fell 7% premarket regardless of unveiling Dragonfly, its new AI data-center model.
-
Qualcomm mentioned it is going to present extra particulars on Dragonfly at its Investor Day on June 24.
-
A lot of the eye at Computex shifted to Nvidia, the place CEO Jensen Huang unveiled the N1X AI PC processor and introduced that Vera Rubin AI chips had entered full manufacturing.
Shares of Qualcomm Inc. (QCOM) slipped 7% in premarket buying and selling on Monday after the chipmaker unveiled its new Dragonfly AI knowledge middle model and outlined its imaginative and prescient for the period of agentic AI, although Nvidia’s PC chip reveal in the end stole a lot of the highlight.
QCOM shares logged their greatest month since September 2019, leaping practically 40% in Could.
Qualcomm’s New AI Knowledge Middle Play
At Taiwan’s Computex convention, Qualcomm introduced the launch of Dragonfly, a brand new model for its AI data-center merchandise. “Meet Qualcomm Dragonfly, our new knowledge middle model. Extra to come back at #QCOMInvestorDay on June 24,” the corporate mentioned in a put up on X after CEO Cristiano Amon’s keynote.
Qualcomm disclosed a couple of further particulars, with Amon reserving contemporary info for its Investor Day on June 24. The transfer marks Qualcomm’s efforts to broaden its attain past smartphones, PCs, and automotive chips and pursue a bigger position within the rising AI infrastructure market, which is at present dominated by Nvidia.
Past Dragonfly, Amon laid out his imaginative and prescient for agentic AI, saying that AI is evolving past assistants and chatbots into autonomous brokers able to working throughout telephones, PCs, automobiles, robots, and industrial programs. Nevertheless, a lot of the eye at Computex in the end shifted towards Nvidia’s entry into AI PCs by means of a brand new processor developed alongside Microsoft and MediaTek.
Qualcomm Bets On Agentic AI
Amon mentioned the digital ecosystem is shifting from being phone-centric to agent-centric, with AI brokers working throughout telephones, PCs, wearables, automobiles and robots. Qualcomm mentioned future units will have to be redesigned for agentic AI, requiring new {hardware} architectures, energy effectivity and deeper integration of sensors and context. The corporate mentioned these brokers will be capable of act on intent, quite than merely reply to prompts.
“The way forward for Agentic AI would require ‘gazillions’ of tokens,” Qualcomm later posted on X, echoing Amon’s remarks through the keynote. He additionally repeated a line he has often used when discussing the adoption of AI: “Resistance is futile.”
