Key Moments
- Qualcomm projects $40 billion in non-handset revenue by fiscal 2029, led by data center, automotive, and industrial/IoT businesses.
- The planned $3.92 billion all-stock acquisition of Modular would give Qualcomm control of the Mojo language and MAX inference engine to reduce GPU ecosystem lock-in.
- Shares have fallen more than 20% over the last 30 days even as options activity showed over 161,000 call contracts traded in a single session and short interest hovered near 2.5%.
Repositioning a Smartphone Supplier as an AI Compute Platform
The semiconductor industry is known for sharp cycles, and Qualcomm (NASDAQ: QCOM) has been caught in the latest downturn. The stock has dropped more than 20% in the last 30 days, pressured by concerns about the smartphone market and supply chain dynamics.
A key factor behind the recent weakness has been near-term memory supply constraints among Chinese handset makers, which has delayed what investors had anticipated as an imminent Android hardware upgrade cycle.
Underneath this volatility, Qualcomm is shifting its business model. The company is building a stack that spans data center and edge artificial intelligence (AI) compute, combining new software assets with server-class silicon anchored by large cloud customers. The strategy is intended to reduce reliance on smartphones and tap into what it sees as a large, long-term opportunity in agentic AI.
Strategic Software Move: Modular Acquisition and the Mojo Stack
Qualcomm is attempting to erode entrenched software moats in enterprise computing, where closed ecosystems have historically made it costly for developers to switch platforms. To that end, the company has agreed to acquire Modular, a software infrastructure provider, in an all-stock transaction valued at $3.92 billion.
The deal, expected to close in the second half of 2026, is designed to bring both intellectual property and key personnel into Qualcomm. That includes Chris Lattner, described as the original architect of several foundational programming languages, including Apple’s Swift.
With Modular under its control, Qualcomm would own the Mojo programming language and the MAX inference engine. This combination is intended to create a silicon-agnostic compute layer, allowing developers to build complex inference applications once and deploy them across diverse hardware environments.
By enabling code to run efficiently on heterogeneous compute, Qualcomm aims to undermine the grip that legacy graphics processing unit providers have maintained over enterprise customers. Lowering these switching barriers is presented as a prerequisite for Qualcomm to gain meaningful share in data center workloads.
| Software Initiative | Details |
|---|---|
| Acquisition target | Modular (software infrastructure firm) |
| Deal value and structure | $3.92 billion, all-stock |
| Expected closing period | Second half of 2026 |
| Key assets | Mojo programming language, MAX inference engine |
Data Center Push: Dragonfly C1000 and Hyperscaler Adoption
Software abstraction alone is not sufficient; Qualcomm is pairing it with new server hardware. At its June 2026 Investor Day, management introduced the Dragonfly C1000 server CPU, based entirely on the custom Oryon architecture.
The Dragonfly C1000 incorporates 250 cores and is described as optimized for agentic AI workloads. Unlike basic generative models that primarily output text, agentic AI is characterized as handling complex, multi-step reasoning and autonomous task execution, requiring sustained, high-throughput compute.
Institutional backing for this architecture emerged quickly. Market data shows that Meta Platforms (NASDAQ: META) entered into a multi-year agreement to use the Dragonfly C1000 in its infrastructure, with shipments set to scale significantly in the second half of 2028.
At the same time, Microsoft Corporation (NASDAQ: MSFT) disclosed plans to deploy Qualcomm’s High Bandwidth Computing architecture within its Azure cloud platform. These relationships are positioned as proof points that the Oryon architecture can support advanced processing demands.
Together, these developments underpin Qualcomm’s new objective of generating $15 billion in data center revenue by fiscal 2029.
| Data Center Initiative | Key Metrics / Partners |
|---|---|
| Dragonfly C1000 CPU | 250-core custom Oryon architecture, focused on agentic AI |
| Meta Platforms (NASDAQ: META) | Multi-year deployment agreement; shipments ramping heavily in second half of 2028 |
| Microsoft (NASDAQ: MSFT) | Adoption of High Bandwidth Computing architecture on Azure |
| Data center revenue goal | $15 billion by fiscal 2029 |




