Amazon and Qualcomm have announced a major long-term partnership to develop customized chips and connectivity technology for the rapidly expanding artificial intelligence data-center market.
Under the multi-generation agreement, Qualcomm will work with Amazon on custom silicon designed primarily for AI inference — the process of running trained AI models to generate responses and perform tasks.
Reuters reports that Amazon could purchase as much as $60 billion worth of Qualcomm AI data-center chips and related products over the life of the partnership.
The agreement also gives Amazon warrants that could allow it to acquire approximately $4 billion worth of Qualcomm shares at $161.26 per share. The warrants will vest as Amazon purchases products under the agreement.
Beyond AI processors, the companies will work together on high-speed optical connectivity technology capable of reaching 1.6 terabits per second, helping data centers move massive amounts of information between AI servers more efficiently.
The partnership represents a major expansion for Qualcomm beyond its traditional smartphone-chip business. The company has been aggressively targeting the AI data-center market as cloud providers search for alternatives and complements to Nvidia’s dominant GPUs.
Amazon joins Microsoft and Meta among major technology companies working with Qualcomm on data-center hardware.
Qualcomm has set a goal of generating around $15 billion in annual data-center chip revenue by 2029, and the Amazon agreement could significantly strengthen its chances of reaching that target.
Amazon, meanwhile, has been rapidly expanding its own custom-chip strategy through AWS as demand for AI infrastructure continues to surge. Its custom silicon business had already reached an annualized revenue run rate exceeding $25 billion by the end of the June quarter.
The two companies are also deepening their relationship in chip development. Qualcomm plans to use more AWS infrastructure and Amazon Bedrock for electronic design automation workloads in an effort to shorten future chip-design cycles.
The deal highlights how competition in AI infrastructure is spreading beyond GPUs into custom processors, networking and optical connectivity as technology companies race to build faster and more energy-efficient data centers.
