Qualcomm and Amazon have announced a multi-generation partnership to develop customised chips for artificial intelligence data centres, with Amazon receiving the right to acquire up to approximately $4 billion in Qualcomm shares.
The agreement, disclosed on September 8, 2026, could involve up to $60 billion in purchases of Qualcomm server chips, technology, systems and manufacturing services over the next decade. The figure represents a potential commercial ceiling linked to the arrangement, not an unconditional order for that amount.
The partnership marks a major expansion of Qualcomm’s ambitions beyond smartphones. It also strengthens Amazon’s effort to develop more of the hardware powering its cloud and AI services rather than relying entirely on chips designed by outside suppliers.
Proposal for Conducting Cyber Crisis Drill, Tabletop Exercise (TTEx) & CCMP Readiness Exercise
Amazon Gets Right to Buy 25 Million Qualcomm Shares
A filing with the US Securities and Exchange Commission shows that Qualcomm issued a warrant to an Amazon affiliate on September 3.
The warrant gives Amazon the right to acquire up to 25 million Qualcomm shares at an exercise price of $161.26 per share. At that price, the full allocation would represent approximately $4.03 billion in share purchases.
The warrant expires on September 3, 2036. Its shares become available in stages tied to commercial agreements, binding purchase orders and actual purchases of Qualcomm products and services.
The filing states that 3.75 million shares vested when the warrant was issued, based on initial purchase commitments. The remaining shares depend on future commercial milestones.
This structure links Amazon’s potential ownership in Qualcomm to the scale of business it conducts with the chipmaker.
What Does a $4 Billion Stock Warrant Mean?
A stock warrant gives its holder the right, but not the obligation, to purchase shares at a specified price before a deadline.
For example, if a company’s share price rises above the warrant’s exercise price, the holder may benefit from the ability to acquire shares at the agreed lower price. If the shares do not become attractive to purchase, the holder is not necessarily required to exercise the warrant.
In this agreement, Amazon’s right to acquire shares is also tied to purchasing Qualcomm products and services. It is therefore an incentive connected to the commercial relationship.
The arrangement should not be described as Amazon immediately investing $4 billion in Qualcomm or as Qualcomm receiving $60 billion in cash. The actual value realised will depend on future purchases, vesting conditions and share-price movements.
New Chips Will Focus on AI Inference
The companies will work together on customised silicon for AI inference, one of the fastest-growing parts of the AI infrastructure market.
AI inference is the process of running a trained artificial intelligence model to generate answers, predictions or other outputs.
Training a model requires enormous computing resources to teach it patterns from data. Inference happens when a user asks a chatbot a question, requests an image or uses an AI-powered service to perform a task.
As millions of people and businesses use AI applications, the cost and speed of inference become increasingly important. Specialised chips can be designed to handle particular workloads more efficiently than general-purpose hardware.
Qualcomm’s expertise in power-efficient processing and chip design could help Amazon develop infrastructure that uses less energy while delivering the performance required for large-scale AI services.
The companies have not disclosed the exact specifications of the jointly developed inference chips or a date when they will be deployed commercially.
Optical Connectivity Is the Other Major Part of the Deal
The partnership is not limited to processors. Qualcomm and Amazon will also develop high-speed optical connectivity solutions for data-centre networks.
The companies said the work will include technologies capable of supporting speeds of up to 1.6 terabits per second, along with future generations.
Modern AI data centres contain large numbers of processors that must exchange information rapidly. If data cannot move between chips and servers quickly enough, expensive computing hardware may sit idle while waiting for information.
Optical connections use light to transmit data and can support high bandwidth across data-centre infrastructure. Qualcomm plans to contribute its serializer-deserializer, or SerDes, and optical digital signal processor technologies.
In simple terms, the project is intended to improve both the computing engines and the high-speed connections that allow those engines to work together.
Qualcomm Expands Beyond Its Smartphone Business
Qualcomm is best known for the Snapdragon processors and communications technology used in smartphones. Its expansion into data-centre hardware is part of an effort to diversify its revenue.
The company has faced pressure from weaker smartphone demand and the gradual reduction of modem-related revenue from Apple, which has been developing its own cellular technology.
Qualcomm has also announced plans to grow its data-centre business and has targeted $15 billion in annual revenue from the segment by fiscal 2029.
The Amazon partnership provides a major cloud customer relationship that could support that ambition. However, the company must still deliver competitive products, secure orders and demonstrate that its technology can meet the demanding requirements of large AI data centres.
The announcement alone does not establish that Qualcomm will achieve its revenue target.
Amazon Builds a Larger Custom Chip Ecosystem
Amazon Web Services already develops several of its own semiconductor families, including Trainium AI accelerators and Graviton processors.
The company’s custom chip strategy is designed to offer customers alternatives to traditional commercial hardware and to improve the economics of its cloud infrastructure.
Working with Qualcomm adds another potential source of specialised technology, particularly for inference and networking.
The arrangement also reflects a wider industry trend in which large cloud companies collaborate directly with chip designers to produce hardware tailored to their own workloads.
That does not mean Amazon will stop buying Nvidia or other suppliers’ chips. Large AI infrastructure providers typically use multiple types of hardware depending on performance, software compatibility, cost and customer requirements.
AWS Tools Will Also Help Qualcomm Design Chips
The partnership is reciprocal. Qualcomm plans to expand its use of AWS infrastructure and artificial intelligence services for electronic design automation workloads.
Electronic design automation, or EDA, refers to the software tools engineers use to design, simulate and verify semiconductor chips before they are manufactured.
Chip design involves testing enormous numbers of components and complex electrical connections. Cloud computing and AI-assisted tools can help engineers run simulations, analyse designs and reduce repetitive work.
Qualcomm said it aims to shorten chip development cycles by using AWS infrastructure, including Amazon Bedrock.
The company has not provided a quantified estimate of the time or cost savings expected from this part of the partnership.
Why the Deal Matters for the AI Chip Market
The agreement comes as technology companies invest heavily in AI data centres, increasing demand for processors, memory, networking equipment and power-efficient infrastructure.
Nvidia remains a dominant supplier of AI accelerators, but cloud companies are increasingly seeking customised alternatives. The demand for inference hardware is also creating opportunities for chipmakers that have traditionally focused on other markets.
Qualcomm’s partnership with Amazon could increase competition and provide cloud operators with additional hardware options. It may also put pressure on established suppliers to improve efficiency and pricing.
For India, where AI infrastructure projects rely heavily on imported advanced semiconductors, greater competition among chip suppliers could eventually affect the cost and availability of computing resources. Any benefit will depend on commercial deployment, pricing and access to the new products.
https://www.linkedin.com/company/policetechnology/
