Denglin Technology
Denglin Technology is a Chinese domestic GPU company founded in 2017 that designs AI inference accelerator cards built on its proprietary GPU+ (MinSky) architecture. It serves enterprise customers across smart city, intelligent computing centers, industrial AI, mining, and agriculture sectors.
- Company typePrivate
- Founded2017
- HeadquartersShanghai, China
- Headcount101–250
- GTM typeB2B
- OfferingHardware or Manufacturing
What Denglin Technology does
Denglin Technology (苏州登临科技股份有限公司) is a privately held Chinese semiconductor company founded in 2017 and headquartered in Suzhou Industrial Park, with additional operations in Shanghai, Beijing, Chengdu, Hangzhou, Shenzhen, and Xi'an. The company designs and manufactures general-purpose GPU (GPGPU) accelerator cards for AI inference workloads, marketed under the Goldwasser (高凛), Nashite (纳适), and KS series brand names. Denglin claims to be China's first domestic GPU company to achieve large-scale commercial deployment, with products sold directly to enterprise customers and through channel and system integration partners.
The company's core technology is its proprietary GPU+ (MinSky) architecture, a software-defined, on-chip heterogeneous design that combines the generality of traditional GPUs with the efficiency of ASICs. The architecture claims 3x better energy efficiency ratio than mainstream international GPU products, supports CUDA and OpenCL ecosystem compatibility, and scales from 70 TOPS in edge inference modules (KS20) to 560 TOPS in datacenter cards (KS38/KS58). The Hamming (瀚铭) software stack provides Day-0 adaptation for major large language models including DeepSeek, GLM, MiniMax M2.5, Kimi, and GPT-OSS, and supports Windows 10, Linux, and Kylin operating systems alongside domestic CPU compatibility. Denglin has filed over 200 domestic and international IP applications covering core IP including chip architecture, microarchitecture, and arithmetic units.
Denglin generates revenue primarily from one-time hardware sales of GPU accelerator products, supplemented by AI platform and solution services bundled with system integration partners. The go-to-market combines direct enterprise sales with a channel partner ecosystem of system integrators and industry-specific solution providers. The company sells across eight industries — internet, smart city, telecom, energy, finance, education, industrial manufacturing, and intelligent computing centers — with named deployments including Shaanxi Coal (mining safety AI), a top Chinese short video platform (content review), an agricultural smart equipment OEM, and Hanten Technology (Shijiazhuang intelligent computing center). Notable channel partnerships include Chengdu Guangshu Huilian Technology, whose AI platform has been deployed to over 1,000 mines across China. Denglin has been recognized as a National Specialized "Little Giant" Enterprise by the Chinese government, a High-Tech Enterprise, and a WAIC 2021 SAIL TOP 30 recipient.
Denglin Technology firmographics
Firmographics- Name
- Denglin Technology
- Legal name
- 苏州登临科技股份有限公司
- Website
- https://denglinai.com
- Company type
- Private
- Founded year
- 2017
- Operating status
- Operating
- Headcount range
- 101–250 employees
- Short description
- Denglin Technology is a Chinese domestic GPU company founded in 2017 that designs AI inference accelerator cards built on its proprietary GPU+ (MinSky) architecture. It serves enterprise customers across smart city, intelligent computing centers, industrial AI, mining, and agriculture sectors.
- Ownership category
- akta.pro rank
Denglin Technology industry classification
Industry- Product category
- AI Accelerator GPUs
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Computer Systems Design Services (541512)
- SIC
- Electronic Computers (3571)
- akta.pro primary industry
- AI Accelerators (GPUs/TPUs/NPUs/ASICs) (HDAAAAAA)
- akta.pro secondary industries
- Compute & Acceleration ICs (CPU/GPU/AI/FPGA/ASIC) (HDAHABAC), GPU/Accelerator Compute Servers (HDABADAC), Graphics Processing Units (GPUs) & Parallel Accelerators (HDAHAJAE)
Keywords
Where Denglin Technology is headquartered
LocationHeadquarters
- HQ city
- Shanghai
- HQ country
- China
- HQ region
- Asia
Offices7 records
Markets served
Denglin Technology business model
Business model- GTM type
- B2B
- Offering type
- Hardware or Manufacturing
- Cost components
- Technology or R&D, Personnel, Supply Chain, Operations, Marketing or Sales, Infrastructure
Revenue model
- GPU Hardware Product Sales: Denglin generates revenue primarily from the sale of proprietary GPU hardware products including the Goldwasser (L series, GS series) and KS series accelerator cards for data center and edge inference deployments. Products are sold directly to enterprise customers and through channel/system integration partners across industries including internet, smart city, telecom, energy, finance, and education. Revenue is one-time hardware purchase based, with potential for follow-on professional services and software support contracts.
- AI Platform & Solution Services: Denglin provides AI platform solutions combining GPU+ hardware with industry-specific AI algorithms, deploying full-chain localized AI systems in partnership with integrators (e.g., 光束慧联 for mining/energy platforms). Revenue includes hardware + platform + service bundled offerings for intelligent computing centers and industry AI deployments.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Other | Multi-year contract | Enterprise hardware products — pricing not publicly disclosed |
Go-to-market motion2 records
Distribution channels2 records
Marketing channels3 records
Denglin Technology product offering
Product offeringCore offering
Denglin Technology designs and manufactures proprietary GPGPU and AI accelerator hardware products (Goldwasser L/GS series and KS series) for data center and edge inference workloads, built on its GPU+ (MinSky) software-defined heterogeneous architecture. The company sells enterprise GPU cards and modules directly to organizations and through system integrators across industries including smart city, internet platforms, mining, industrial AI, intelligent computing centers, education, and energy. It also provides the Hamming software stack with day-0 support for major large language models (DeepSeek, GLM, MiniMax, Kimi, GPT-OSS) and CUDA/OpenCL compatibility.
Product overview
Denglin Technology (苏州登临科技股份有限公司) offers a comprehensive GPU product portfolio built on its proprietary GPU+ (MinSky) architecture. The product lineup spans three generations: Goldwasser I L series (first-generation datacenter cards), Goldwasser II GS series (second-generation datacenter GPUs), and KS series (third-generation datacenter products and edge inference boards). The Goldwasser I UL/MXM series provides low-power edge inference modules, while the KS edge inference products (KS19/KS20) target high-concurrency video scenarios. All products are supported by the Hamming® software stack and are compatible with CUDA/OpenCL ecosystems, enabling seamless algorithm migration. The company targets smart cities, intelligent computing centers, industry large models, AIPC, industrial vision, internet, education, energy, and retail sectors.
Differentiator
Problem solved
Functional benefit
Brands
- Goldwasser (高凛): Proprietary general-purpose GPU product line for data center and edge computing applications
- Nashite (纳适)
- Hamming (瀚铭)
- GPU+
Products and services
- Goldwasser I L Series (L128/L256)
- Goldwasser II GS Series (GS30/GS40) Second-generation datacenter GPUs with 32-128GB memory per card and up to 192GB/s bandwidth, optimized for education, smart city, and enterprise private cloud general large models and high-concurrency inference.
- KS Series Datacenter Products (KS28/KS38) Third-generation datacenter products for AI server high-density deployment, supporting FP32 to INT4 precision with 32-128GB memory, up to 409.6GB/s bandwidth, and efficient multi-card scheduling for large models and system integration.
- Goldwasser I UL/MXM Series (UL32/UL64/MXM32/MXM64) Low-power edge inference modules with <20W power consumption and 4-64GB memory, designed for compact edge deployments in education and smart city scenarios including AI proctoring and real-time video analysis.
- KS Series Edge Inference Products (KS19/KS20) Second-generation edge inference boards with 8-64GB memory and 25W typical power consumption, supporting up to 100 concurrent video streams per card for multi-classroom and compact AI server integration.
- NashII (纳适II) Series Accelerator Cards Industrial-grade accelerator cards providing 70-560 TOPS effective AI computing power and 8-128GB memory, supporting high-resolution video parsing, complex behavior recognition, and multi-target classification across edge to cloud deployments.
Quantifiable outcome
- KS28-based content review system achieves 80-channel 1080P real-time video review per card, 99.2% accuracy, and 65% power reduction vs prior solutions
- +7 more outcomes
Companies that use Denglin Technology
Customer profileNamed customers4 records
Segments8 records
Ideal customer profiles5 records
Denglin Technology technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability16 records
Feature5 records
Denglin Technology partnerships and signals
Strategic signalScale indicators5 records
Recent moves6 records
Expansion highlights6 records
Denglin Technology competitors and assessment
Company assessmentBroad incumbents
- NVIDIA: Global GPU/AI accelerator leader whose CUDA software ecosystem and H100/A100-class data center GPUs define the performance and developer-experience benchmark that Denglin's GPU+ chips and Hamming stack explicitly aim to match.
- Huawei (Ascend): Established Chinese incumbent with the Ascend line of AI training/inference accelerators and a vertically integrated chip-to-cloud stack; dominates the domestic-GPU procurement landscape and constrains Denglin's share.
Direct peers
- Cambricon (寒武纪): Listed Chinese AI chip designer covering cloud training/inference accelerators and edge SoCs; sets the listed-comp benchmark for domestic Chinese AI accelerators and competes head-on with Denglin for sovereign compute customers.
- MetaX (沐曦集成电路): Domestic Chinese AI chip startup developing high-performance GPUs for data center AI training and inference; repeatedly cited alongside Denglin as a top-tier Chinese GPU contender.
- Moore Threads: Chinese fabless GPU designer building full-stack general-purpose GPUs for AI, graphics and cloud; competes with Denglin in the same domestic-GPU opportunity and targets inference workloads via CUDA-compatible stacks.
- Iluvatar CoreX (天数智芯): Chinese AI GPU vendor producing cloud training and inference accelerators; directly comparable to Denglin in targeting domestic data-center AI compute and offering CUDA-compatible software.
- Moffett AI (墨芯人工智能): Early-stage Chinese AI inference chip designer; co-listed with Denglin as a leading domestic AI accelerator and competes for the same data-center inference workloads.
- Enflame (燧原科技): Chinese AI accelerator startup producing cloud training and inference chips; directly listed with Denglin as a leading domestic Chinese AI-chip vendor and shares the same CUDA-equivalent software compatibility narrative.
- Biren Technology (壁仞科技): Chinese GPU startup developing general-purpose and AI training chips; competes for the same domestic cloud/AI compute customers as Denglin with a similar fabless chip plus software-stack story.
Regional players
- Hygon Information Technology (海光信息): Chinese-listed x86-compatible compute and AI accelerator vendor serving domestic data centers; comparable as a domestically designed high-performance compute provider catering to Chinese AI infrastructure demand, adjacent rather than a direct GPU rival.
Market position
Strengths4 records
Weaknesses4 records
Competitive moat5 records
Key risks5 records
Key highlights6 records
Customer concentration
Denglin Technology social profiles
Digital presenceDenglin Technology financial estimates
Financial estimateRevenue estimate
Valuation estimate
Denglin Technology leadership team
Management profileNumber of profiles
Profiles2 records
Denglin Technology funding detail
Funding detailFunding overview
Funding rounds4 records
Investors8 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Denglin Technology M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Denglin Technology
What does Denglin Technology do?
Denglin Technology designs and manufactures proprietary GPGPU and AI accelerator hardware products (Goldwasser L/GS series and KS series) for data center and edge inference workloads, built on its GPU+ (MinSky) software-defined heterogeneous architecture. The company sells enterprise GPU cards and modules directly to organizations and through system integrators across industries including smart city, internet platforms, mining, industrial AI, intelligent computing centers, education, and energy. It also provides the Hamming software stack with day-0 support for major large language models (DeepSeek, GLM, MiniMax, Kimi, GPT-OSS) and CUDA/OpenCL compatibility.
Is Denglin Technology a public or private company?
Denglin Technology is a private company. It is classified as venture growth investor backed and is currently operating.
When was Denglin Technology founded?
Denglin Technology was founded in 2017. It employs 101 to 250 people.
Where is Denglin Technology based?
Denglin Technology is headquartered in Shanghai, China, in the Asia region.
How does Denglin Technology make money?
Two revenue lines are on record. GPU Hardware Product Sales are the primary driver. The others are AI Platform & Solution Services.
Who are Denglin Technology's main competitors?
Broad incumbents on record are NVIDIA and Huawei (Ascend). Direct peers are Cambricon (寒武纪), MetaX (沐曦集成电路), Moore Threads, Iluvatar CoreX (天数智芯), Moffett AI (墨芯人工智能), Enflame (燧原科技) and Biren Technology (壁仞科技). Hygon Information Technology (海光信息) is listed as a regional player.
Does Denglin Technology have an API?
No public API is recorded for Denglin Technology.
What industry is Denglin Technology in?
Denglin Technology's product category is AI Accelerator GPUs. Its primary akta.pro industry code is HDAAAAAA, AI Accelerators (GPUs/TPUs/NPUs/ASICs), with a secondary code of HDAHABAC, Compute & Acceleration ICs (CPU/GPU/AI/FPGA/ASIC). Its NAICS code is 51821 and its SIC code is 3571.