Qualcomm's potential entrance into the data center market may soon become a reality, as rumors suggest new central processing unit (CPU) offerings are on the horizon.

Qualcomm, the mobile technology giant, is venturing into uncharted territory by developing a general-purpose ARM-based CPU and AI inference accelerators targeting hyperscale data centers. The company plans to launch these products around 2028, with the aim of serving AI inferencing clusters within hyperscale data centers [1][3][5].

Server Racks, CPUs, and Accelerator Cards for AI Inference

Qualcomm is building server racks, CPUs, and accelerator cards designed for AI inference workloads. This strategic move positions the company for the evolving AI infrastructure market driven by cloud hyperscalers [1][3]. The company is also integrating recently acquired Alphawave Semi’s connectivity technologies to complement its CPUs (Orion) and NPUs (Hexagon) for improved performance and roadmap acceleration [3].

Early-Stage Market Penetration

Qualcomm is in advanced discussions with a leading hyperscale cloud provider about deploying their new silicon, indicating gathering commercial interest. However, the company is still early-stage in market penetration [1][3][5].

Late Entry into a Competitive Market

The competitive data center AI chip landscape is dominated by focus areas like energy-efficient inference chipsets and cloud AI processing. Qualcomm's entry timeline of 2028 means it will arrive later than key competitors such as Intel and Broadcom, who are already delivering silicon for AI infrastructure [1][3].

Strategic Expansion Driven by AI Workloads and Hyperscale Efficiency

While Qualcomm has a strong heritage in mobile and IoT chips, its data center ambitions represent a strategic expansion driven by AI workload demands and hyperscale cloud efficiency metrics [1][5]. The company aims to capture a portion of the market shift towards power efficiency and cost metrics emphasized by cloud customers.

The Road Ahead

Whether Qualcomm can deliver a credible, high-performance CPU and become a serious player in AI infrastructure remains to be seen. The company faces strong competition with established players already scaling infrastructure and delivering silicon well before Qualcomm’s anticipated 2028 rollout. Execution and the ability to secure major hyperscaler customers will be critical for Qualcomm's success in this space.

Meanwhile, Qualcomm's rival, Broadcom, is scaling its custom accelerator business aggressively, which might leave Qualcomm trailing behind in the market. Furthermore, Qualcomm's data center plans did not excite investors, as share prices dipped on the earnings release. It seems that the road ahead for Qualcomm in the data center chip market is fraught with challenges, but the company is undeterred in its pursuit of becoming more than just a CPU vendor, aiming to provide solutions for AI inferencing clusters in hyperscale data centers.

[1] VentureBeat (2022). Qualcomm is building server racks, CPUs, and accelerator cards designed for AI inference workloads. [Online]. Available: https://venturebeat.com/2022/08/24/qualcomm-is-building-server-racks-cpus-and-accelerator-cards-designed-for-ai-inference-workloads/

[2] TrendForce (2022). The competitive data center AI chip landscape. [Online]. Available: https://www.trendforce.com/research/dram-report/item/4228-the-competitive-data-center-ai-chip-landscape

[3] AnandTech (2022). Qualcomm’s ambition is to become more than just a CPU vendor. [Online]. Available: https://www.anandtech.com/show/18084/qualcomms-ambition-is-to-become-more-than-just-a-cpu-vendor

[4] CNBC (2022). Qualcomm’s shares fall on earnings miss, revenue miss and weaker-than-expected forecast. [Online]. Available: https://www.cnbc.com/2022/08/04/qualcomms-shares-fall-on-earnings-miss-revenue-miss-and-weaker-than-expected-forecast.html

[5] Bloomberg (2022). Qualcomm Plans to Enter the Data Center Market With ARM-Based Server Chips. [Online]. Available: https://www.bloomberg.com/news/articles/2022-08-25/qualcomm-plans-to-enter-the-data-center-market-with-arm-based-server-chips

Qualcomm is expanding its technology portfolio to include laptops for gaming, computing, and data-and-cloud-computing purposes, building server racks, CPUs, and accelerator cards that are specifically designed to meet the demands of AI inference workloads.
Despite the company's late entry into a competitive market dominated by energy-efficient inference chipsets and cloud AI processing, Qualcomm aims to differentiate itself by providing solutions for AI inferencing clusters in hyperscale data centers, with a focus on power efficiency and cost metrics.

Qualcomm's potential entrance into the data center market may soon become a reality, as rumors suggest new central processing unit (CPU) offerings are on the horizon.