Qualcomm's $60 Billion Deal With Amazon
Qualcomm (QCOM) is building chips for Amazon (AMZN) under a new partnership announced this morning. The deal carries a headline value of $60 billion in potential business, with about a $9 billion commitment already in hand based on the warrant structure. If it succeeds, $60 billion is not the ceiling.
The work spans custom silicon and networking. Qualcomm's (QCOM) strength in low-power chips gives Amazon (AMZN) a new tool set that sits alongside its own homegrown chips and the merchant silicon it already buys. The deal helps Qualcomm hit the $15 billion target it has talked about for some time. It should quiet the skeptics who doubted Qualcomm could enter this space, at least for now.
For years there has been talk about Qualcomm's (QCOM) data center business and whether the company could become a real force there. As this market grew from hundreds of billions to trillions of dollars in total addressable market (TAM), Qualcomm needed a few anchor customers and wins to prove itself. This is one.
From Handsets to AI
Qualcomm (QCOM) was long tied tightly to handsets. When Apple (AAPL) moved away from Qualcomm's chip business, some investors doubted the company could fully recover. It did. Qualcomm then moved into automotive and IoT. The market still felt Qualcomm had not taken its technology and made the most of what is now a multi-trillion dollar TAM. Intel (INTC) faced some of this same view.
The cumulative TAM for AI capital spending between now and 2030 runs up to $14 trillion. Qualcomm was not in that space, so it made smart acquisitions across software, networking, and intellectual property. That lets it build custom chips to speed up AI workloads, plus CPUs, competing with Intel (INTC), AMD (AMD), ARM (ARM), and others. Qualcomm can make both custom and off-the-shelf CPUs for AI.
Amazon (AMZN) is the customer being talked about now, but this is likely the start of several such partnerships for Qualcomm, all powered by AI. Qualcomm has already made multiple deals. Meta Platforms (META) is set to use the Dragonwing (referred to as "Dragonfly") C1000 processing units, and there is work with Microsoft's (MSFT) Azure.
Why AWS Needs More Partners
Amazon Web Services (AWS) is a bright spot for Amazon (AMZN) and the largest hyperscale cloud by a wide margin. Early in the AI boom, AWS had to play catch-up while Google (GOOGL) and Microsoft (MSFT) moved fast. Amazon had the largest footprint of all cloud providers and made heavy investments, building its own CPU chips (Graviton) and its own custom AI chips (Trainium, referred to as "Tranium").
Over the last few quarters, AWS growth has accelerated in a meaningful way and the business has caught up. Now Amazon (AMZN) wants all the compute it can get. Nobody has enough compute. Amazon cannot build enough, cannot buy enough Nvidia (NVDA), cannot buy enough AMD (AMD), and cannot make enough of its own chips.
Some asked whether this deal signals strength for Qualcomm (QCOM) and weakness for AWS. The answer: it is strength across the board. Amazon (AMZN) simply cannot build enough compute, so it is hunting for new, highly capable partners, and Qualcomm showed up at the right time.
The market often wants to read these deals as "less Nvidia" or "less Graviton and Trainium." The better read is that it all adds up together. Forecasts for AI capex and AI compute keep growing so fast that it is an all-hands-on-deck situation. Qualcomm's (QCOM) history, its low-power design, and its acquisitions fit well with what Amazon (AMZN) is trying to do. Amazon is racing to ramp compute to meet the several hundred billion in backlog sitting in its cloud business today.
IonQ and Quantum
IonQ (IONQ) held its investor day, with the stock up 1.5%. It raised full-year guidance on the back of the SkyWater acquisition, which adds foundry capabilities. IonQ's strategy is vertical integration across quantum computing, quantum networking, and quantum sensing. Its guidance now puts it at $100 million a quarter in revenue this year, with meaningful growth ahead.
Quantum has been a very speculative space, but IonQ (IONQ) has been executing. More government investment showed up today in some of its peers. The national security angle makes quantum more important, and government backing gives the whole field more validation across capabilities from computing to foundry. Expect much more about quantum in 2027. Quantum is what comes next after AI, and the two will move together for the long haul.
