Qualcomm Lands AWS Partnership to Build AI Data Centre Chips, Shares Jump 10%
The smartphone chip giant is making a serious push into the data centre market, and Amazon just gave it a very public vote of confidence.
Key points
- Qualcomm shares rose 10% on Tuesday after the company announced a data centre chip partnership with Amazon Web Services.
- The deal focuses on inference, the process by which an AI model answers a question or completes a task after it has already been trained.
- Qualcomm unveiled its first data centre processor, the Dragonfly C1000, in June 2026, with Meta already committed to using it from 2028.
- Qualcomm is targeting $15 billion in data centre revenue by its 2029 financial year.
- Both Meta and Amazon spend hundreds of billions of dollars per year building AI infrastructure, making either company a significant customer.
Qualcomm, best known as the company whose chips power most Android smartphones, is making a determined move into artificial intelligence data centres. The company's shares jumped 10% on Tuesday after it announced a partnership with Amazon Web Services, or AWS, the cloud computing arm of Amazon that hosts vast amounts of the world's internet infrastructure.
The two companies said they will work together across multiple generations of custom silicon, meaning chips designed specifically for Amazon's needs, to build out AWS's AI infrastructure. The focus is on inference: the moment an AI model actually does its job, producing an answer, a translation, or a piece of code for a user.
Why does this matter to ordinary people?
Anyone who uses an AI assistant, a shopping recommendation, or a voice search is on the receiving end of inference. Faster, cheaper inference means snappier AI tools that cost less to run, savings that can eventually reach consumers.
Qualcomm's pitch is power efficiency. Data centres consume enormous amounts of electricity, and a chip that delivers strong performance while using less power is a genuine selling point, both for the companies paying the energy bills and for the environment.
The company made its data centre ambitions official in June 2026, when it unveiled a processor called the Dragonfly C1000. The chip is built for agentic AI, meaning AI software that can carry out multi-step tasks on its own rather than simply answering a single question. Meta, the parent company of Facebook and Instagram, said it would put the Dragonfly C1000 into production from 2028.
How does this stack up against Qualcomm's rivals?
Nvidia currently dominates the AI chip market. Its graphics processors, the specialised chips that do the heavy number-crunching AI needs, have become the default hardware for training and running AI models. Qualcomm is not yet a direct rival at that scale, but the AWS deal is its second major hyperscaler endorsement in months. Hyperscalers are the handful of technology giants, including Amazon, Google, and Microsoft, that each spend tens of billions of dollars building and running cloud infrastructure.
Qualcomm has set itself a target of $15 billion in data centre sales by its 2029 financial year. The company also has a roadmap that includes a dedicated AI accelerator chip and a product designed to link multiple chips together for heavier workloads.
| Milestone | Detail |
|---|---|
| Dragonfly C1000 announced | June 2026 |
| Meta production commitment | From 2028 |
| AWS partnership announced | June 2026 |
| Data centre revenue target | $15 billion by FY2029 |
For now, the AWS deal is a statement of intent as much as a product launch. No shipment volumes or financial terms were disclosed. The partnership signals that at least two of the world's biggest AI spenders see Qualcomm as a credible alternative worth investing in, which is itself meaningful in a market Nvidia has dominated for years.
Common questions
What is inference, and why does it matter?
Inference is what happens when an AI model responds to you: the trained model reads your input and produces an output. It is distinct from training, which is the far more expensive process of teaching the model in the first place. Most of the day-to-day cost of running AI services comes from inference, so chips that do it efficiently save money at scale.
Does this affect the AI tools I use today?
Not immediately. The Qualcomm chips tied to this partnership will not reach production until 2028 at the earliest. Longer term, wider competition in the AI chip market could help bring down the cost of cloud AI services.
Is Qualcomm actually competing with Nvidia yet?
Not at scale, not yet. Nvidia holds a commanding position in AI chips, built over years of software support as well as hardware performance. Qualcomm is entering the market with a focus on power efficiency and custom designs for specific customers, a different angle rather than a direct head-to-head.
