Amazon is plowing ahead with its custom-silicon strategy, announcing a major “multi-generational product collaboration” with Qualcomm Technologies. In case you need a refresher, Qualcomm is a major U.S. technology company known for designing semiconductors and wireless communication technologies.
Under the new agreement terms, Amazon could make up to $60 billion in payments for Qualcomm server-chip products, technology, systems, and manufacturing services. Amazon also received a warrant giving it the right to buy up to 25 million Qualcomm shares, which at close to $162 per share could work out to roughly $4 billion.
How Will the Collaboration Work in Practice?
“This collaboration with Qualcomm Technologies builds on a strong foundation of partnership and reflects our shared commitment to pushing the boundaries of what’s possible,” said Prasad Kalyanaraman, Vice President, AWS. “By working together on customized silicon and advanced connectivity, we’re delivering more performant, efficient, and cost-effective infrastructure for our customers.”
So, what exactly’s on the table? Essentially, the deal combines AWS’s expertise in AI infrastructure with Qualcomm’s silicon-design capabilities. The two companies will work across multiple generations of customized silicon to support AI inference in large-scale data centers.
The collaboration will also see the two companies work together on an optical connectivity solution to support the growing scale and demands of this AI infrastructure. The new solutions will tap Qualcomm’s SerDes and optical DSP technologies to bundle and ship data from chip to chip, with solutions extending to speeds of up to 1.6 Tbps.
Beyond this, Qualcomm plans to deepen its use of AWS AI infrastructure, including Amazon Bedrock, for electronic design automation (EDA) workloads, with the aim of reducing chip-design cycles.
Final Thoughts
What Amazon is shooting for here is greater control and flexibility over its AI infrastructure stack at a time when demand for data center capacity is spiraling. AWS already has a custom-silicon strategy, including its Graviton CPUs, Trainium AI chips, and Inferentia chips.
However, making these deeper investments in custom silicon could help AWS tailor hardware more closely to specific workloads and potentially reduce its reliance on third parties, which is always a huge bonus, regardless of industry or era.
While it’s still too early to say whether the collaboration with Qualcomm will lead to lower costs or enhanced performance, that’s the aim here. And there’s another angle worth noting. The collaboration with Qualcomm doesn’t mean AWS is moving away from NVIDIA.
In fact, AWS and NVIDIA just expanded their partnership in August 2026, announcing plans to deploy 2 million additional NVIDIA GPUs across AWS’s global infrastructure in 2027 and 2028. The companies are also deepening their work across AI factories, CPUs, networking, open models, data processing and robotics.
NVIDIA GPUs remain an important part of AWS’s AI infrastructure, even as Amazon continues developing its own custom chips. What the Qualcomm collaboration does is add another layer to its strategy, giving AWS access to additional customized silicon. AWS is following what is now a well-trodden path in the AI Era, and that’s the multi-vendor route, combining NVIDIA GPUs, its own custom chips and customized silicon developed with Qualcomm rather than relying on a single chip supplier.
