Qualcomm launches Dragonfly C1000 CPU for data centers with Meta as key client

Qualcomm introduced its new central processing unit for data centers, the Dragonfly C1000, on Wednesday and disclosed that Meta plans to adopt the processor once volume manufacturing begins in 2028.

Qualcomm targets data center expansion beyond smartphones. Executives detailed the chip at an investor event, describing it as engineered specifically for agentic AI workloads while delivering strong computing output with restrained energy demands.

The company laid out plans to pursue the rapidly expanding data center sector through multiple offerings. Among them are a dedicated AI processor and technology to link several chips into unified systems. CEO Cristiano Amon noted that Qualcomm has assembled relevant technologies and now holds a full suite of solutions ready for the market's next stage.

Meta deal builds on existing hyperscaler relationships. CFO Akash Palkhiwala told investors the firm already serves nearly every major hyperscaler through its mobile and other legacy lines. He characterized the Meta agreement as an extension of prior edge successes, bolstered by Qualcomm's manufacturing reach, technical depth and established trust. "This is not a new relationship," Palkhiwala said.

Investor interest grows in CPUs for AI agents. The disclosure arrives while analysts anticipate that central processors will assume greater shares of tasks traditionally handled by graphics processors and specialized AI accelerators, driven by the rise of autonomous AI agents. Palkhiwala observed that CPU supply remains constrained and that additional vendors will be required.

Qualcomm shares declined during Wednesday's trading session.

Smartphone revenue still dominates but diversification accelerates. Mobile devices generated roughly two-thirds of product revenue in the quarter ending in March. The firm is broadening its focus toward automotive, robotics and data-center applications, segments that analysts project will outpace smartphones, whose global shipments topped out in 2017.

Qualcomm executives highlighted how its experience designing power-efficient chips for phones and laptops aligns with the needs of hyperscalers, whose facilities increasingly face electricity constraints. The company also reported winning two contracts to develop custom silicon for such cloud operators.

Acquisition bolsters AI software capabilities. In a separate move, Qualcomm purchased Modular for an undisclosed sum. The startup created tools allowing AI programs to operate across varied processor types; Qualcomm positions the technology as comparable to Nvidia's CUDA platform widely used in AI development.