Partnership strengthens Qualcomm’s push into AI data centers as it seeks to diversify beyond smartphones and challenge Nvidia’s dominance in accelerated computing
Qualcomm said Tuesday that Amazon could purchase as much as $60 billion of its artificial intelligence data-center chips and related products under a long-term partnership, giving the chipmaker a major potential customer as it accelerates its expansion beyond smartphones.
The agreement marks a significant vote of confidence in Qualcomm’s emerging data-center business and could help the company establish itself as a major supplier of AI infrastructure alongside Nvidia and other semiconductor companies competing for spending from cloud providers.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Qualcomm will also grant Amazon warrants valued at about $4 billion that vest as the cloud company purchases Qualcomm products, according to a regulatory filing. The warrants give Amazon the right to buy Qualcomm shares at $161.26 each.
Qualcomm shares rose more than 3% following the announcement.
The deal comes as Qualcomm works to reduce its reliance on the smartphone market, where growth has weakened and component costs have increased. The company also faces the eventual loss of its modem business with Apple, making the development of new revenue streams important.
For the past year, Qualcomm has been courting major cloud companies with custom AI processors and other data-center technologies as hyperscalers seek alternatives to Nvidia’s dominant AI accelerators.
Amazon now joins Microsoft and Meta among major technology companies supporting Qualcomm’s push. Qualcomm expects its data-center chip business to generate $15 billion in annual revenue by 2029.
“The deal is exactly the kind of development that Qualcomm needed to reassure the market that the lofty data-center ambitions they set for themselves could indeed be met,” said Bob O’Donnell, chief analyst at TECHnalysis Research.
The partnership will initially focus on chips designed for AI inference, the stage of AI computing in which trained models are deployed to generate responses, predictions and other outputs.
Inference is becoming an important semiconductor market as AI applications move from training large models toward widespread commercial deployment. That shift could create opportunities for specialized chips designed to deliver high performance at lower power and cost than general-purpose accelerators.
The agreement also underscores Amazon’s growing role as both a customer and developer of custom AI silicon. Amazon Web Services’ custom-chip business had an annualized revenue run rate of more than $25 billion at the end of the June quarter, making internally developed processors an increasingly important part of the cloud division’s infrastructure strategy.
Amazon has developed its own AI chips as it seeks greater control over costs and performance and to reduce dependence on external suppliers. Qualcomm’s involvement gives the company another potential source of specialized computing capacity.
The deal follows a similar arrangement between Marvell Technology and Google’s parent company, Alphabet, announced weeks ago, under which Google received the right to acquire a stake in Marvell worth as much as $12.2 billion.
The increasingly common structure indicates that the AI infrastructure race is creating closer financial and commercial ties between chipmakers and hyperscalers. Cloud companies want assured access to computing capacity and customized silicon, while semiconductor suppliers are seeking long-term commitments that can justify the enormous investment required to develop AI processors.
From Chips to Optical Connectivity
Qualcomm’s agreement with Amazon extends beyond computing processors. The companies will also develop optical communications technology for AI data centers, including high-speed connectivity solutions capable of supporting speeds of up to 1.6 terabits per second.
The move gives Qualcomm exposure to another critical bottleneck in AI infrastructure. As data centers deploy more powerful processors, moving enormous volumes of data between chips, servers and storage systems has become nearly as important as computing power itself.
Qualcomm’s entry into optical connectivity is relatively recent. The company acquired AlphaWave for $2.4 billion last year, and AlphaWave CEO Tony Pialis subsequently became Qualcomm’s head of data-center chips.
“By incorporating both compute and optical interconnect, Qualcomm is highlighting the range of semiconductor technologies that they can uniquely bring to AI infrastructure,” O’Donnell said.
The strategy could give Qualcomm a broader role in data-center architectures rather than limiting it to supplying individual processors. Qualcomm also plans to expand its use of AWS services and infrastructure for chip-design workloads under the partnership, with the aim of shortening development cycles.
The agreement means more than a large potential sales contract to Qualcomm. It is seen as an important test of whether the company can turn its years of expertise in low-power mobile processors into a meaningful position in the rapidly expanding data-center market.
Nvidia remains the dominant supplier of AI accelerators, but hyperscalers have increasingly sought customized chips to lower costs, improve energy efficiency and gain greater control over their infrastructure. Google, Amazon and Microsoft have all invested heavily in developing their own silicon, while companies such as Qualcomm and Marvell are positioning themselves as partners in that transition.
The potential $60 billion commitment from Amazon would therefore give Qualcomm significant visibility as it builds out a business that is still small relative to its established handset operations. At the same time, the warrant structure aligns Amazon’s interests with Qualcomm’s ability to deliver. The warrants vest as Amazon purchases products, effectively tying the potential equity upside to commercial adoption.
The challenge will be converting the headline potential of the agreement into sustained revenue and profitable market share. AI infrastructure is capital-intensive, competition is escalating rapidly, and hyperscalers have considerable bargaining power because they can develop chips internally or source them from multiple suppliers.
Still, securing Amazon as a potential customer at this scale materially strengthens Qualcomm’s case that its diversification into AI infrastructure can become a significant new growth engine just as its traditional smartphone business faces structural challenges.