Qualcomm is building a long AI hardware pipeline. On September 8, 2026, the company announced a multi-generational product collaboration with Amazon to supply customized silicon at scale for large-scale AI data centers, with both companies working together on AI inference.
The collaboration also covers optical connectivity solutions extending up to 1.6T, along with future-generation connectivity products. Qualcomm will use its SerDes and optical DSP technologies to support high-bandwidth interconnects across Amazon’s data center networks.
That matters because AI workloads are not only demanding more compute. They are also putting pressure on the systems that move data between processors, memory, and the wider network. Qualcomm said the collaboration is driven by those infrastructure demands, pairing Amazon’s AI infrastructure with Qualcomm’s power-efficient processing, silicon design, and system-level integration.
A roadmap built around inference
Qualcomm has already placed several pieces of that roadmap on the table. At its June 24, 2026 Investor Day, the company introduced the Dragonfly C1000 CPU, the Dragonfly AI300 inference accelerator, High Bandwidth Compute technology, connectivity products, and custom silicon solutions.
The C1000 is a purpose-built data center CPU built around custom-designed Oryon cores in a chiplet design. It includes more than 250 cores, with commercial availability expected in 2028, and Qualcomm said it is estimated to deliver more than twice the performance.
The AI300 is also expected to begin commercial sampling in 2028. Qualcomm designed it for inference and said it can deliver four to eight times better performance per watt than existing GPU-based architectures on memory bandwidth per watt per card.
Before that arrives, High Bandwidth Compute Gen 1 with the AI250 is expected to sample in mid-2027. The sequence gives Qualcomm a staged path from connectivity and custom silicon to inference hardware, rather than one grand product launch followed by the usual long silence.
Amazon becomes both customer and infrastructure partner
Qualcomm also plans to deepen its own use of AWS AI infrastructure, including Amazon Bedrock, for electronic design automation workloads. That creates a two-way relationship: Qualcomm supplies customized silicon and connectivity for AWS data centers while using AWS infrastructure in its own chip-design work.
Prasad Kalyanaraman, Vice President at AWS, said the collaboration builds on an existing partnership between the two companies. He said the joint work on customized silicon and advanced connectivity aims to deliver more performant, efficient, and cost-effective infrastructure for AWS customers.
Cristiano Amon, President and CEO of Qualcomm, said, “As AI demand accelerates, data center infrastructure will require advances in both computing and connectivity to deliver greater performance with more efficiency.” The statement neatly captures the deal’s central point: faster processors alone will not solve the infrastructure problem if the network cannot keep up.
ASML is also part of the wider AI chip manufacturing ecosystem, with plans to work with major chipmakers to use advanced tools for larger chips. Qualcomm’s collaboration with Amazon sits inside that broader shift, where chip design, manufacturing tools, data center systems, and optical links must advance together.
The commercial dates remain ahead: mid-2027 for High Bandwidth Compute Gen 1 with the AI250, and 2028 for both the C1000’s expected availability and the AI300’s expected commercial sampling. For AWS, the announcement adds a future stream of custom processors and connectivity options. For Qualcomm, it turns its data center ambitions into a multi-generational relationship with one of the biggest infrastructure platforms in the market.
Based on
