Qualcomm secures Meta as customer for new AI data center processor

Qualcomm just landed one of the biggest names in tech as a customer for its nascent data center chip business. Meta will use the company’s Dragonfly C1000 processor in its data centers, giving Qualcomm’s ambitious push beyond smartphones a serious credibility boost.

There’s a catch, though. The chip won’t be available until 2028, which means Qualcomm is selling a vision of the future while competitors like Nvidia and AMD are already shipping silicon by the truckload. Wall Street noticed: QCOM shares dropped nearly 7% in premarket trading following the broader Dragonfly brand announcement at COMPUTEX 2026.

Qualcomm’s Dragonfly play

The Dragonfly brand is Qualcomm’s new umbrella for everything data center. Think server CPUs, AI accelerators, and custom ASICs, all designed to help the company diversify away from its heavy reliance on mobile chips. Qualcomm unveiled the brand at COMPUTEX on June 1-2, and CEO Cristiano Amon pointed to ongoing collaborations with major hyperscalers as proof the strategy is gaining traction.

The Meta deal for the Dragonfly C1000 sits within a broader lineup that includes the AI200, aimed for 2026, and the AI250, targeting 2027. In other words, Qualcomm isn’t just building one chip and hoping for the best. It’s constructing an entire product roadmap designed to compete across the data center AI stack.

And Meta isn’t Qualcomm’s only hyperscaler win. In May 2026, the company secured a deal with ByteDance to supply several million AI ASICs for autonomous AI agents.

Why the stock dropped anyway

Landing Meta as a customer sounds like the kind of news that should send a stock soaring. So why did shares fall nearly 7% in premarket trading?

The simplest explanation: investors were hoping for something more immediate. A chip that ships in 2028 is still two years away from generating revenue. Qualcomm’s ByteDance deal, with its commitment of several million ASICs, provides some concrete revenue grounding. But the Meta partnership, with its 2028 timeline, reads more like a letter of intent than a purchase order.

Qualcomm plans to offer more detail during its Investor Day on June 24, 2026.

What this means for the AI chip landscape

Meta is simultaneously developing in-house MTIA chips and signing up for Qualcomm’s Dragonfly C1000, which tells you something interesting: even companies with custom chip programs recognize they can’t build everything themselves.

Qualcomm’s pitch centers on AI inference, the process of running trained models rather than training them from scratch. Qualcomm’s mobile heritage, built on squeezing maximum performance out of minimal power budgets, maps neatly onto inference workloads that favor efficiency and cost-per-query over raw compute throughput.

Qualcomm’s Investor Day on June 24 should reveal how the company plans to address the software side of this equation. For investors, the key metrics to track are concrete revenue projections from the ByteDance ASIC deal and any additional hyperscaler commitments that shrink the gap between announcement and actual chip shipment.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.