Key Findings
Qualcomm is expanding its Dragonfly Connectivity Platform, integrating AI accelerators, custom silicon, and networking capabilities, to empower next-generation AI factories. The company’s roadmap includes 1.6T optical modules (O200) and Active Electrical Cables (AECs) (CU200) for 2026-2027, and 3.2T optical modules (O400) and AECs (CU400) by 2028, all designed to support the high-bandwidth connectivity essential for demanding AI workloads.
Technical and Market Details
Qualcomm’s expanded Dragonfly Connectivity Platform features new coherent-lite optical solutions for scale-out applications (up to 20 km) and PAM4 Optical SerDes (Serializer/Deserializer) for shorter distances (up to 2 km). The coherent-lite optical solutions optimize signal quality and power efficiency for long-distance data transmission, enabling inter-data center connectivity and distributed AI clusters. PAM4 Optical SerDes facilitates high-density data transmission over shorter distances, making it suitable for integration with in-packaging solutions like Co-Packaged Optics (CPO) and Near-Package Optics (NPO).
The company’s roadmap is designed to keep pace with the rapid evolution of bandwidth requirements in AI data centers. The 1.6T optical modules will address the primary demands of current AI training and inference workloads, while the 3.2T optical modules are engineered to meet the needs of even larger and more complex AI models in the future. These optical modules and AECs will integrate tightly with Qualcomm’s custom AI chips and Ethernet switching solutions, providing a comprehensive ‘one-stop’ computing platform for AI data centers.
Background and Industry Context
As the scale of generative AI models expands exponentially, bandwidth requirements for both internal and external data center connectivity are increasing exponentially. Traditional electrical interconnects are increasingly struggling to handle this immense data volume efficiently and with low latency, leading to a severe ‘data bottleneck.’ The focus of chip manufacturers like Qualcomm on not only AI accelerators but also the underlying connectivity technologies highlights the importance of system-level optimization in the AI era. Optical technology is an indispensable component for resolving this bottleneck, and Qualcomm’s comprehensive approach aims to deliver the scalability and performance required by AI factories.
Strategic Significance and Outlook
Qualcomm’s expansion of its Dragonfly Connectivity Platform and its ambitious optical module roadmap will significantly impact the construction of next-generation AI factories. The introduction of 1.6T and 3.2T optical modules will substantially improve the efficiency of AI training and enable the development and deployment of larger and more complex AI models. By integrating optical technology with AI accelerators, custom silicon, and networking, Qualcomm is strengthening its competitiveness in the AI infrastructure market and is expected to play a crucial role in shaping the future of AI. This initiative will accelerate ‘Moore’s Law for AI’ from the perspective of connectivity technology.
Source: https://wccftech.com/qualcomm-dragonfly/
Get our weekly technology intelligence — free
Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.
Subscribe Free — Weekly Tech Intelligence
By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.
- Your email and selected fields are used only to deliver the newsletter.
- We never share your information with third parties.
- You can unsubscribe anytime via the link in each email.
See our Privacy Policy for details.
Takes about a minute · Unsubscribe anytime
Comments