Qualcomm is deepening its push into artificial intelligence infrastructure through a multi-generation partnership with Amazon Web Services (AWS), putting greater focus on the competitive landscape for custom AI chips and data-centre connectivity.
The companies announced plans to collaborate on custom silicon for AI inference as well as high-speed optical connectivity for data centres. The agreement represents an important step for Qualcomm as it seeks to expand beyond its traditional strength in smartphone processors and modems.
The development could also have implications for other semiconductor companies, including MediaTek and Alchip Technologies, which operate in areas connected to custom chip design and advanced computing infrastructure.
Qualcomm and Amazon Expand Custom AI Chip Partnership
Under the agreement, Amazon could purchase up to $60 billion of Qualcomm’s AI data-centre chips and related products over the course of the long-term arrangement, according to the supplied report.
Amazon will also receive approximately $4 billion worth of Qualcomm warrants, allowing it to purchase Qualcomm shares at $161.26 per share. The warrants are structured to vest as Amazon purchases Qualcomm products.
The companies plan to develop customised AI chips primarily aimed at AI inference, the stage at which trained AI models process new data and generate responses or predictions.
Their collaboration will also cover high-speed optical connectivity, an increasingly important component of large-scale data-centre infrastructure as AI workloads drive demand for faster movement of data between computing systems.
Qualcomm Wants a Bigger Role in AI Infrastructure
For Qualcomm, the Amazon relationship strengthens its effort to establish itself as a data-centre semiconductor supplier.
The company has set a target of reaching $15 billion in annual data-centre chip revenue by 2029. Securing a major hyperscaler as a customer provides a significant commercial opportunity, although the potential purchase figure should not be treated as guaranteed revenue.
Amazon joins other major technology companies that have been working with Qualcomm on customised AI silicon, potentially giving the chipmaker a broader customer base outside smartphones.
The timing is also significant for Qualcomm because its handset business faces structural challenges, including the eventual loss of Apple’s modem business. Expanding into data centres could provide the company with another source of higher-value semiconductor revenue.
Why AI Inference Matters
Much of the early AI infrastructure boom has focused on training increasingly large models. However, inference becomes increasingly important once AI systems are deployed at scale.
Every time an AI model generates an answer, recommendation, image or other output, computing resources are required for inference.
That creates an opportunity for chipmakers capable of delivering efficient inference hardware. Qualcomm’s strategy is therefore not limited to competing for traditional AI training workloads.
The company’s optical-connectivity initiative could further broaden its addressable market as data centres require increasingly sophisticated networking infrastructure to connect large numbers of processors.
What the Amazon Deal Means for AWS
Amazon has already developed a substantial custom-chip business through AWS, including its own silicon for cloud workloads.
The partnership with Qualcomm could give AWS another source of customised computing technology and diversify its AI infrastructure options.
For Amazon, one potential advantage of custom silicon is the ability to optimise hardware around particular workloads, potentially improving the balance between performance and cost.
Qualcomm’s emphasis on inference is particularly relevant as AI applications become more widely deployed across cloud services.
The combination of custom computing and optical connectivity could also allow AWS to explore more tightly integrated approaches to data-centre design.
The Competitive Impact on MediaTek and Alchip
Qualcomm’s expansion into AI data-centre silicon creates a broader competitive issue for semiconductor companies pursuing similar opportunities.
MediaTek has been expanding beyond smartphones into areas such as high-performance computing, connectivity and data-centre-related technologies. A larger Qualcomm presence in custom AI infrastructure could increase competitive pressure as major technology companies look for alternatives to established AI-chip suppliers.
Alchip Technologies, meanwhile, is known for its role in designing and developing customised application-specific integrated circuits (ASICs). The growing demand for hyperscaler-designed AI chips has created opportunities for companies involved in custom silicon development.
Qualcomm’s direct relationship with Amazon potentially strengthens its position in that ecosystem because it combines custom chip development with its own semiconductor technology and connectivity capabilities.
However, the competitive effect will depend on the chips Qualcomm ultimately develops and whether they meet Amazon’s performance, efficiency and cost requirements.
The $60 Billion Figure Comes With Important Conditions
The headline purchase potential should not be interpreted as $60 billion in guaranteed Qualcomm revenue.
Amazon’s purchases are tied to a long-term arrangement, and Qualcomm must still develop and commercialise chips that meet the requirements of a highly competitive AI semiconductor market.
Nvidia remains a dominant force in AI computing, while hyperscalers including Amazon are simultaneously developing their own silicon to reduce costs and optimise hardware for specific workloads.
Qualcomm therefore faces the challenge of competing not only with traditional chipmakers but also with customers that increasingly possess significant internal chip-design capabilities.
Amazon Also Faces Capital Spending Pressure
The partnership offers Amazon additional technology options, but it comes as the company continues to invest heavily in AI infrastructure.
Building data centres capable of supporting large-scale AI workloads requires substantial spending on processors, networking equipment, power systems, cooling and physical infrastructure.
Amazon’s commitment to Qualcomm therefore needs to be viewed alongside its broader AI capital expenditure.
The supplied report also notes that Amazon was preparing a sterling bond sale as major technology companies increasingly seek financing for the enormous cost of expanding AI infrastructure.
If AI services generate strong demand and attractive returns, the spending could support long-term growth. If returns disappoint, however, the scale of infrastructure investment could place greater pressure on free cash flow and capital efficiency.
Qualcomm’s $4 Billion Warrants Add Another Consideration
The warrants awarded to Amazon are another important element of the agreement.
Amazon’s ability to purchase Qualcomm shares at $161.26 could give it an additional economic interest in the success of the partnership. For existing Qualcomm shareholders, however, the potential exercise of warrants could result in dilution.
The ultimate effect will depend on Qualcomm’s future share price, the vesting conditions and whether Amazon exercises the warrants.
What Investors Should Watch
The Qualcomm-Amazon agreement creates a potentially significant new avenue for both companies, but execution will determine its eventual financial impact.
Investors tracking Qualcomm should pay particular attention to:
- Progress toward its $15 billion annual data-centre revenue target for 2029
- Commercial deployment of Qualcomm’s custom AI inference chips
- Adoption by Amazon and other hyperscalers
- AI chip performance and power efficiency
- Growth in Qualcomm’s optical-connectivity business
- Competition from Nvidia and other AI semiconductor companies
- Potential dilution associated with Amazon’s warrants
For Amazon, the key issues include AI infrastructure spending, AWS growth, custom-chip adoption and the returns generated from its expanding AI investments.
The Bigger Picture
Qualcomm’s agreement with Amazon marks a significant expansion of its ambitions in AI infrastructure. The company’s traditional smartphone semiconductor business remains important, but data-centre chips and high-speed connectivity could become increasingly relevant to its long-term growth strategy.
For AWS, Qualcomm adds another potential source of custom AI silicon at a time when hyperscalers are seeking greater control over computing costs and performance.
The deal could also increase competitive pressure across the custom-chip ecosystem, including companies such as MediaTek and Alchip. But its ultimate significance will depend on execution: Qualcomm must turn the partnership into commercially competitive products, while Amazon must determine whether those products deliver the performance and economics required by its rapidly expanding AI infrastructure.
