On Wednesday, Qualcomm (QCOM) delivered a highly aggressive growth signal at its New York Investor Day, announcing a major push into the AI data center chip arena and sharply raising its long-term financial targets. Boosted by the news, Qualcomm’s shares rallied significantly in after-hours trading.
Qualcomm projects that by fiscal 2029, its data center business will generate more than $15 billion in annual revenue, while overall revenue from businesses outside smartphones will reach $40 billion—nearly doubling the $22 billion target set two years ago. The company also expects that by then, handset chip revenue will account for only one-third of total chip business revenue. Chief Financial Officer Akash Palkhiwala made it clear that the company will achieve true business diversification.
On the specific growth path, the data center business target is clear and aggressive. Palkhiwala revealed that in fiscal 2027, which begins in October this year, the business will contribute $5 billion in revenue, with $1 billion coming from newly expanded custom chip customers. By fiscal 2029, the automotive chip revenue target has also been raised to $10 billion, with the company stating that its “automotive design-win pipeline” has expanded to $65 billion. Qualcomm also expects adjusted earnings per share to exceed $18 in fiscal 2029, far surpassing analysts’ estimate of $15.26.
To achieve these goals, Qualcomm unveiled a series of major products and partnerships, formally entering direct competition in the data center space with giants such as Nvidia (NVDA) and Broadcom (AVGO).
Qualcomm introduced the “Dragonfly C1000,” a central processor designed specifically for AI data centers, and announced that Meta Platforms (META) has agreed to adopt the chip in its infrastructure. This CPU will enter mass production in 2028, and Meta will also adopt subsequent iterations of the product. Qualcomm emphasized that the processor is built for “agentic AI,” focusing on delivering high-performance computing power at lower power consumption.
At the same time, Microsoft (MSFT) has also become a customer for its new category of AI chips. The chip leverages a technology path Qualcomm calls “high-bandwidth computing,” using low-cost memory chips commonly found in smartphones and laptops to replace the expensive high-bandwidth memory used by Nvidia and the SRAM memory used by Cerebras (CBRS). Tony Pialis, head of Qualcomm’s data center business, said this brings tremendous price-performance advantages to the industry.
In addition to the publicly named giants, Qualcomm also revealed that it has won custom chip contracts from two other unnamed “hyperscale” cloud service providers, with related revenue expected to begin generating before the end of this year. Pialis described the process by saying, “Customers pulled us in, rather than us forcing our way in.”
To Better Compete With Nvidia’s CUDA Ecosystem, Qualcomm announced on the same day that it has reached an agreement to acquire AI software startup Modular Inc. for approximately $3.9 billion in stock. The company develops software that enables AI applications to run on multiple chip architectures, and is seen as filling the software capability gap Qualcomm needs to enter the data center market.
The market responded positively to Qualcomm’s ambitious vision. In regular trading on Wednesday, Qualcomm shares closed at $197.41, up 16% for the year. Following the release of the optimistic outlook in after-hours trading, the stock surged again, with the maximum gain reaching more than 16%. Even Arm (ARM), which provides underlying technology for many of Qualcomm’s chips, saw its shares rise as much as 8%.
