Qualcomm Unveils AI Chips, Sets $15B Target in AI

Qualcomm unveiled data center AI chips and revealed its CPU customer is Meta. The company aims to capture $15 billion in AI chip sales, competing with Nvidia.

QCOM made its most ambitious datacenter move yet on June 24, unveiling the Dragonfly C1000 CPU and naming META as anchor customer for a multi-generation server deployment targeting agentic AI workloads. The chip is also planned for Microsoft Azure deployment in mid-2027, with a custom high-bandwidth compute architecture designed to undercut NVDA's CUDA ecosystem lock-in. Qualcomm shares jumped 15% in extended trading following the announcement.

The $15B AI chip revenue target is staged across a clear roadmap: $5 billion by fiscal 2027 scaling to $15 billion by fiscal 2029. Qualcomm simultaneously raised its total non-handset revenue target to $40 billion by 2029, more than doubling the prior $22B projection. The full-stack approach includes custom CPUs, AI accelerators, memory technology, and a software acquisition aimed at enabling developers to exit the CUDA ecosystem without significant re-tooling costs.

The announcement represents a structural inflection for Qualcomm, long categorized as a mobile-chip company. QCOM's path to datacenter credibility runs through Meta's willingness to stake its next-generation server fleet on an unproven datacenter silicon provider. That design win carries significant third-party validation weight for other hyperscalers evaluating Nvidia alternatives. Investors are watching whether Qualcomm can convert the Meta relationship into broader wins before Nvidia tightens its ecosystem with Blackwell Ultra later in 2026.