Qualcomm has signed a long-term deal that lets Amazon purchase up to $60 billion of its AI data-center chips and related products. The agreement, announced Tuesday, includes warrants worth about $4 billion that allow Amazon to buy Qualcomm shares at $161.26 apiece as product purchases vest. Shares of Qualcomm rose more than 3% on the news.
The partnership marks a major step in Qualcomm's push to diversify beyond smartphones. The company faces the eventual loss of its Apple modem business, rising component costs, and softer handset demand. Over the past year, it has courted cloud providers with custom AI chips and data-center technology as those companies search for alternatives to Nvidia's dominant processors. Amazon now joins Microsoft and Meta as customers backing that effort. Qualcomm expects data-center chip revenue to reach $15 billion 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.
AI inference takes center stage
Under the agreement, Qualcomm and Amazon will collaborate on chips for AI inference - the process of running trained AI models in production. This segment has become a key battleground among semiconductor firms as enterprises shift from training models to deploying them at scale. The work aligns with broader trends in Generative AI and LLM deployment, where inference efficiency directly impacts cost and latency.
Amazon's custom chip business has already become a growth driver for its cloud unit, with an annualized revenue run-rate above $25 billion at the end of the June quarter. The Qualcomm partnership adds another layer to that strategy, coming weeks after Marvell Technology struck a similar custom AI chip deal with Alphabet's Google that included the right to buy a stake worth up to $12.2 billion.
Optical connectivity and chip design
Beyond computing silicon, the deal includes optical communications technology from Qualcomm. The companies plan to develop high-speed optical connectivity solutions extending to 1.6 terabits per second, addressing the bandwidth demands of AI data centers. This marks a vote of confidence for a business Qualcomm entered through its $2.4 billion acquisition of AlphaWave last year - a move that brought AlphaWave CEO Tony Pialis in as Qualcomm's data-center chip chief.
"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.
Qualcomm also plans to expand its use of Amazon Web Services for chip design workloads, aiming to shorten development cycles. The arrangement mirrors a growing pattern of intertwined financing in the AI boom, where chip deals come bundled with equity components and cloud commitments.
Why this matters for IT and development professionals
For teams managing AI for IT & Development workloads, the Qualcomm-Amazon deal signals an accelerating shift in the infrastructure market. More chip suppliers competing for cloud provider business means more options for inference hardware - and potentially lower costs over time. The optical connectivity work, targeting 1.6 terabits per second, addresses a real bottleneck: as AI clusters scale, network bandwidth often becomes the limiting factor before compute does. Developers and architects planning multi-year infrastructure roadmaps should track how these custom silicon partnerships reshape the pricing and availability of cloud AI resources.
Your membership also unlocks:
