Partnerships

Qualcomm to Supply AWS With Custom AI Chips Across Multiple Generations

mm
Published
Add Unite.AI to your preferred sources on Google

Qualcomm said on September 8, 2026 that it has entered a multi-generational product collaboration with Amazon to supply customized silicon at scale for large-scale AI data centers, with the two companies working together on AI inference.

The announcement from Qualcomm covers multiple generations of customized silicon intended to support Amazon Web Services’ AI infrastructure. Beyond the silicon work, the companies said they are collaborating on optical connectivity solutions extending up to 1.6T, along with future-generation connectivity solutions. Qualcomm said the connectivity effort will draw on its SerDes and optical DSP technologies to support high-bandwidth interconnects within Amazon’s data center networks.

Qualcomm framed the agreement around the demands that growing AI workloads place on infrastructure. As those workloads expand, the company said, they drive demand for compute, storage, networking, memory bandwidth, and energy-efficient infrastructure, and the collaboration pairs Amazon’s AI infrastructure with Qualcomm’s power-efficient processing, silicon design, and system-level integration.

The deal also has a reverse component. Qualcomm said it plans to deepen its own use of AWS AI infrastructure, including Amazon Bedrock, for electronic design automation workloads, targeting a reduction in chip design cycles.

What the Companies Said

“As AI demand accelerates, data center infrastructure will require advances in both computing and connectivity to deliver greater performance with more efficiency,” said Cristiano Amon, President and CEO of Qualcomm Incorporated, adding that Qualcomm brings decades of work in advanced processing and power-efficient compute to the collaboration.

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 is aimed at delivering more performant, efficient, and cost-effective infrastructure for AWS customers.

Qualcomm’s Data Center Push

The Amazon agreement lands three months after Qualcomm laid out a broader data center portfolio at its June 24, 2026 Investor Day in New York. At that event, the company introduced the Dragonfly C1000 CPU, the Dragonfly AI300 inference accelerator, its High Bandwidth Compute technology, connectivity products, and custom silicon solutions. The AI300 joined the previously announced AI200 and AI250 in what Qualcomm described as a multi-generation AI accelerator roadmap on an annual cadence.

Qualcomm said the C1000 is a purpose-built data center CPU using custom-designed Oryon cores in a chiplet design with more than 250 cores, with commercial availability expected in 2028. The company said the AI300, its third-generation rack-level AI inference platform, is expected to begin commercial sampling in 2028, and that its High Bandwidth Compute Gen 1 with the AI250 is expected to sample in mid-2027. Qualcomm said the AI300 is designed to deliver four to eight times better performance per watt than existing GPU-based architectures on memory bandwidth per watt per card, and that the C1000 is estimated to deliver more than twice the performance per watt of existing server CPU benchmarks based on specifications. These are Qualcomm’s own design targets and estimates.

At the same June event, Qualcomm announced a multi-year, multi-generation agreement with Meta, under which the Dragonfly C1000 is planned to power Meta’s next-generation server fleet. The company said more than 35 organizations across the technology and AI ecosystems expressed support for its data center vision at the time.

Connectivity Portfolio

Qualcomm’s June roadmap also detailed a connectivity portfolio spanning die-to-die, copper, optical, and campus-reach interconnects, with support for 800G and 1.6T connectivity across optical, active optical cable, and active electrical cable applications at reaches up to 20 kilometers. The company said that portfolio combines its SerDes, PAM4, coherent-lite DSP, signal integrity, and telemetry capabilities. The Amazon collaboration applies that optical connectivity work, including the 1.6T generation, to Amazon’s data center networks.

The companies did not disclose financial terms, a deployment timeline, or volume commitments for the customized silicon in the announcement.

mm

Theo Nash is an AI-generated specialist at Unite.AI, covering AI infrastructure, compute, and the hardware systems that power modern artificial intelligence. His work focuses on the technical foundations behind large-scale AI workloads, including data centers, accelerators, networking, and the software stacks that tie them together.

With an analytical and engineering-driven perspective, Theo examines how advances in GPUs, custom silicon, memory architectures, and distributed systems enable new generations of AI models. He pays particular attention to performance trade-offs, energy efficiency, scalability, and the practical constraints that shape real-world deployment of AI infrastructure.

Articles authored by Theo Nash are AI-generated and reviewed by Unite.AI’s editorial team to ensure technical accuracy, clarity, and responsible coverage of the rapidly evolving AI compute landscape.