
Qualcomm Dragonfly™ products redefine data centers with industry-leading performance and power efficiency to achieve sustained and scalable AI infrastructure growth.
Qualcomm branded products are products of Qualcomm Technologies, Inc. and/or its subsidiaries.

Benefits
With 768 GB per card and 36.8 TB per rack, AI200 serves models from 7B up to 10 trillion parameters, supporting long-context, RAG, and agentic workloads without compromise.
Purpose-built for inference economics, AI200 delivers compelling performance-per-dollar, driving down cost-per-token for generative and agentic AI workloads.
The Qualcomm® AI Inference Suite provides end-to-end tooling for model deployment as bare-metal, cloud VM, or inference-as-a-service, reducing complexity and accelerating time-to-value.
A fully integrated OCP ORv3-compliant, direct-liquid-cooled rack with cableless backplane simplifies deployment, accelerates time-to-production, and improves reliability across enterprise and hyperscale data centers.
Use Cases
From foundation models to enterprise AI applications, Qualcomm Dragonfly provides infrastructure to build, test, and scale what’s next.
Serve and scale foundation models.
Support multi-agent reasoning workflows.
Assistants, copilots, and support.
Enterprise knowledge search and Q&A.
Summarization and extraction workflows.
Business automation and AI copilots.
Qualcomm Dragonfly™
Memory bandwidth and energy are the AI inference infrastructure bottlenecks. Qualcomm Dragonfly™ solutions are engineered for leading bandwidth, tokens-per-watt, and tokens-per-dollar.
Tokens / (second * watt)
up to 8x
better than GPU based systems¹
Memory bandwidth / watt
6x
higher than HBM based systems²
Memory capacity / watt
200x
better than SRAM based systems³
Performance / watt
>2x
better than the latest server CPU systems⁴
1 With Qualcomm HBC; Qualcomm estimates compared to contemporary GPU-based architectures on decode performance for select models. 2 With Qualcomm HBC; Qualcomm estimates compared to competing published product specifications normalized at card-level. 3 With Qualcomm HBC; Qualcomm estimates compared to competing published product specifications normalized at rack-level. 4 With Qualcomm Dragonfly C1000 CPU; Qualcomm estimates compared to existing product benchmarks for server CPU competitive offerings based on specs.
Qualcomm branded products are products of Qualcomm Technologies, Inc. and/or its subsidiaries.

Rack-scale inference platform with superior memory capacity and low TCO.
Learn MoreQualcomm Dragonfly AI200 is a purpose-built, rack-scale AI inference accelerator designed to support generative AI and agentic AI workloads.
768 GB
LPDDR memory per accelerator
36.8 TB
Memory capacity per rack
355 TB/s
Memory bandwidth per rack
Liquid-cooled
ORv3-compliant rack design
Why Cirrascale
Beyond infrastructure, you get a true partner. Our hands-on team sets you up and keeps systems running smoothly, reducing your infrastructure burden.
Every organization is different. We customize deployments to your resources, funding model, and team, so your setup grows and scales with you.
From hardware to software, we deliver a full-stack solution built for performance, reliability, and scale, so you spend less time on integration and more on results.
Get Started
Explore AI inference powered by Qualcomm managed by Cirrascale, with access to purpose-built infrastructure, software resources, and deployment options for testing and production workloads.