Artificial Intelligence 12 min read

DPU Industry Overview: Applications, Competitive Landscape, and Market Outlook

The 2021 China DPU Industry Development White Paper outlines DPU’s key use cases across data centers, cloud, security, AI, edge, storage and streaming, analyzes the competitive landscape among major vendors, and projects rapid global and Chinese market growth through 2025.

Architects' Tech Alliance
Architects' Tech Alliance
Architects' Tech Alliance
DPU Industry Overview: Applications, Competitive Landscape, and Market Outlook

DPU (Data Processing Unit) is an emerging hardware accelerator designed to offload networking, security, storage, and compute tasks from CPUs in data centers and cloud environments, with typical scenarios including data center and cloud computing, network security, high‑performance computing and AI, telecom/edge computing, data storage, and streaming media.

The white paper details six major application areas: (1) data center and cloud (physical, virtual, containerized, private, public, hybrid clouds); (2) network security (distributed security, NGFW, micro‑segmentation); (3) HPC and AI (cloud‑native supercomputing, deep‑learning recommendation acceleration); (4) telecom and edge (telecom cloud, edge computing); (5) storage (hyper‑converged, elastic block, instance storage); (6) streaming (high‑quality visual, 8K video, CDN).

The DPU market is described as a blue‑sea with many competitors. Major players such as Intel, NVIDIA, AWS, Alibaba Cloud, and Huawei are developing their own solutions. Three architectural approaches are identified: (a) general‑purpose many‑core processors (e.g., Broadcom Stingray); (b) specialized heterogeneous core arrays (e.g., IPU); (c) hybrid designs that combine programmable cores with dedicated accelerators (e.g., NVIDIA BlueField‑3, Fungible).

Company‑specific analyses include: NVIDIA DPU – built on Mellanox networking and ARM cores, focusing on security, networking, and storage offload; Intel IPU – a programmable infrastructure processor that accelerates security, virtualization, storage, and load‑balancing; Xilinx – Smart NIC based on FPGA with optional ARM cores for offloading IPSec, ML, DPI, and video transcoding; Marvell OCTEON 10 – 5 nm ARM Neoverse N2‑based DPU with PCIe 5.0, DDR5, and vector packet processing; Zhongke Yushu – KPU‑based DPU offering network protocol, database, big‑data, storage, and encryption acceleration.

Globally, DPU technology is still in its early stage, requiring ecosystem development. Intel, Marvell, and NVIDIA are identified as the most promising pioneers, each pursuing a distinct integration of general‑purpose cores, FPGA, and ASIC accelerators. The worldwide DPU market was $30.5 billion in 2020 and is projected to exceed $245.3 billion by 2025.

In China, the DPU sector is nascent but rapidly growing, driven by massive internet traffic, heightened security needs, and AI workloads that CPUs and GPUs cannot efficiently handle. The Chinese DPU market reached ¥3.9 billion in 2020 and is expected to surpass ¥565.9 billion by 2025.

The full analysis, including detailed charts and references, is available in the 2021 China DPU Industry Development White Paper.

cloud computingNetwork Securitymarket analysisAI accelerationdata centersemiconductorDPU
Architects' Tech Alliance
Written by

Architects' Tech Alliance

Sharing project experiences, insights into cutting-edge architectures, focusing on cloud computing, microservices, big data, hyper-convergence, storage, data protection, artificial intelligence, industry practices and solutions.

0 followers
Reader feedback

How this landed with the community

login Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.