What DataCanvas does

DataCanvas (九章云极) is a Beijing-headquartered AI cloud and compute infrastructure provider operating under the Alaya NeW (九章智算云) brand. The company delivers a platform-plus-modules stack centered on the Alaya NeW self-serve cloud, offering GPU compute instances (A Series agile, E Series advanced, B Series 10,000-card flagship), elastic and dedicated container clusters, cloud container instances, dedicated Slurm clusters, the JiuZhang Intelligent Computing OS for cross-cluster scheduling, and the AI Factory automated compute-to-delivery pipeline. Customers purchase compute via the standardized 'du' (度) unit (1 du = 312 TFLOPS * 1H, priced at 18 yuan/du), subscription packages for AI coding and OpenClaw AI assistant, or enterprise platform services; the company claims an average 36% cost reduction versus traditional GPU compute and GPU cluster utilization exceeding 65%.

The platform supports Day0 adaptation for frontier models such as DeepSeek V4 and GLM-5.2, and is integrated with mainstream DevOps tooling (kubectl, Helm, CI/CD GitOps). DataCanvas operates 30,000P of compute across 10+ intelligent computing centers in 10+ regions, serving 200+ enterprise and team customers spanning manufacturing, pharma R&D, cultural tourism, steel, shipping, e-commerce, multimodal AI startups, and academic research. Distribution is hybrid: a product-led self-serve motion on alayanew.com plus enterprise direct sales, with a sovereign-level international expansion underway via a tripartite MoU signed in January 2026 with Sharjah Communications Technology Authority and AI Caravan to develop data center infrastructure in the UAE.

Founded in 2013 and led by Chairman/CEO/founder Lei Fang and CTO Mindong Shang Shang, DataCanvas has raised approximately $120M+ across nine disclosed rounds through 2025, including late-2023 capital from China Electronics Smart Fund and a late-2025 round led by Beijing Artificial Intelligence Industry Investment Fund and Beijing Information Industry Development Investment Fund. The company remains private, headquartered in Beijing, and is actively executing on domestic vertical expansion, international geography, and recurring-revenue product extensions.

Name

DataCanvas

Legal name

DataCanvas

Website

https://zetyun.com

Company type

Private

Founded year

2013

Operating status

Operating

Headcount range

11–50 employees

Short description

DataCanvas (九章云极) is a Beijing-based AI cloud infrastructure provider operating the Alaya NeW platform, delivering GPU compute instances, elastic container clusters, the JiuZhang Intelligent Computing OS, and AI Factory automation to 200+ enterprise customers across manufacturing, shipping, pharma, research, and AI development segments.

Ownership category

akta.pro rank

Product category

AI Cloud Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518210), Computer Systems Design and Related Services (54151), Software Publishers (5132)

SIC

Services-Computer Programming, Data Processing, Etc. (7370), Services-Computer Integrated Systems Design (7373), Services-Computer Processing & Data Preparation (7374)

akta.pro primary industry

AI Compute Cloud & GPU-as-a-Service (HDAAAAAK)

akta.pro secondary industries

AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers) (HDAAAAAG), Compute Resource Management & Scheduling (Cluster/Workload Managers) (HDABADAK), GPU-Accelerated & AI Training/Inference Servers (HDACABAG)

Keywords

• GPU cloud computing
• AI infrastructure platform
• Intelligent computing services
• Large model inference
• Containerized AI workloads

Headquarters

HQ city

Beijing

HQ country

China

HQ region

Asia

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Operations

Revenue model

1. GPU Compute Resource Sales: Self-serve cloud platform selling GPU computing resources on a per-unit basis (18 yuan per 'du') and subscription packages for AI coding and dedicated assistants. Customers can purchase A-series, E-series, and B-series GPU instances.
2. AI Factory and Platform Services: AI factory solutions, large model development kits, model inference platforms, and intelligent computing center monitoring provided as platform services to enterprise customers.

Pricing tiers

Go-to-market motion1 record

Distribution channels1 record

Marketing channels3 records

Core offering

DataCanvas sells GPU compute resources and AI development infrastructure through its Alaya NeW intelligent computing cloud platform, including A/E/B series GPU instances, cloud container instances, elastic and dedicated container clusters, large model inference and training platforms, the AI Factory automation pipeline, OpenClaw 24/7 AI assistant, and the JiuZhang Intelligent Computing OS for cross-cluster scheduling. Offerings are priced on standardized "du" units, pay-as-you-go, or subscription, and target AI developers, researchers, and enterprise customers running large model training and inference workloads.

Product overview

DataCanvas International operates its cloud AI infrastructure platform under the Alaya NeW (九章智算云) brand. The product portfolio is structured as a platform-plus-modules architecture centered on the Alaya NeW cloud platform, which provides GPU compute instances, container clusters, and intelligent scheduling. The offering includes: the Alaya Lab one-stop AI development platform and AI Factory for automated model training-to-delivery pipelines; cloud container instances (CCI), elastic container clusters, and dedicated container/Slurm clusters for GPU workloads at various scales; the intelligent computing software platform covering elastic container platform (VKS), large model development toolkit, and dedicated container platform (DKS); and inference and training tools including OpenClaw 24/7 AI assistant, DeepSeek V4, and GLM-5.2 flagship models. The infrastructure layer comprises the 九章智算操作系统 (Alaya Operating System) for cross-cluster scheduling and unified monitoring, plus physical infrastructure including all-flash/AI storage and high-performance computing networks. GPU instances are available in A Series (agile/development), E Series (advanced/training), and B Series (benchmark/peak) tiers. The platform also offers the 「任我享」 flexible compute package for on-demand GPU usage.

Differentiator

Problem solved

Functional benefit

Products and services

• Alaya NeW Intelligent Computing Cloud (九章智算云) Core AI cloud platform providing GPU compute instances, container clusters, AI development tools, and intelligent scheduling for AI model training and inference workloads, sold via self-serve portal at alayanew.com and through enterprise consultation.
• AI Factory (AI 工厂) Enterprise solution automating the pipeline from computing power production to intelligent delivery, defining intelligent industrialization of AI compute infrastructure for AI training and inference workloads.
• Alaya Lab (九章智算实验室) One-stop AI development platform providing high-end compute so developers can quickly deploy and iterate AI models and experiments.
• OpenClaw AI Assistant (开放爪) 7x24 intelligent AI assistant with one-click deployment of dedicated AI assistants for enterprises, available on the Alaya NeW platform.
• Cloud Container Instances (云容器实例, CCI) Containerized GPU cloud instances providing stable compute power with real-time monitoring, on-demand billing, and container lifecycle management for enterprises with containerized workloads.
• Elastic Container Cluster (弹性容器集群) High-performance GPU elastic container cluster with resource auto-scaling based on business load, supporting kubectl, Helm tools and CI/CD GitOps workflows.
• Dedicated Container Cluster (专属容器集群) Dedicated GPU container cluster providing higher performance guarantees and exclusive GPU resources for demanding enterprise workloads.
• Dedicated Slurm Cluster (专属 Slurm 集群) Managed Slurm resource scheduling cluster for high-performance computing workloads.
• Intelligent Computing Software Platform (智算软件平台) Comprehensive software platform including elastic container platform (VKS), large model development toolkit, and dedicated container platform (DKS) for model training and inference.
• Large Model Inference Platform (大模型推理平台) Platform for deploying and running large AI model inference workloads with optimized performance.
• GPU Instances - A Series, E Series, B Series (GPU 实例) Three-tier GPU compute lineup: A Series (Agile) for rapid model development validation with low threshold and Serverless architecture, E Series (High-end) for complex model training and high-bandwidth computing with large video memory, and B Series (Flagship) supporting ten-thousand-card scale and peak throughput for extreme intelligence computing ceilings.
• "Ren Wo Xiang" Basic Compute Package (「任我享」基础算力包) Flexible pay-as-you-go GPU compute package priced at 18 yuan per "du" unit, based on Serverless architecture with second-level task startup and zero idle cost - charging only for effective computation time, with no subscription obligation, unlimited time and no card limit.
• Alaya Operating System (九章智算操作系统) Operating system layer providing cross-cluster computing power elastic scheduling, serverless elasticity, and integrated computing network for unified intelligent computing center monitoring.
• AI Coding Subscription Package (Alaya Token) Subscription package combining powerful compute with mainstream AI models and mainstream AI programming tools so customers can code with AI enhancement on demand.

Quantifiable outcome

• Average cost reduction of 36% compared to traditional GPU compute approaches
• +3 more outcomes

Named customers9 records

Segments7 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability5 records

Feature10 records

Partnerships

Three partnerships are on record, tiered core.

• Sharjah Communications Technology Authority (SCTA)coreStrategic or Co-development Partner · 23 June 2026Advanced from MoU stage (January 2026) to active implementation planning phase. SCTA CEO led delegation to DataCanvas Beijing headquarters to explore project procedures, tour AI data center operations, and align on milestones and timelines for joint AI infrastructure initiative in the Emirate of Sharjah. Closed-door meetings established a clear roadmap and defined responsibilities.
• Sharjah Communications Technology Authority (SCTA)coreStrategic or Co-development Partner · 27 January 2026Tripartite MoU signed with DataCanvas International and AI Caravan to collaborate on developing data center infrastructure and AI solutions in the Emirate of Sharjah, UAE. Witnessed by Sheikh Sultan bin Ahmed bin Sultan Al Qasimi, Deputy Ruler of Sharjah. All three parties committed to jointly prepare business plans, exchange technical expertise, and position Sharjah as a competitive data center hub in the Middle East. Implementation planning phase began June 2026 with delegation visiting Beijing headquarters.
• AI CaravancoreStrategic or Co-development Partner · 27 January 2026Tripartite MoU with DataCanvas International and Sharjah Communications Technology Authority to develop advanced data centers and modernize AI technology infrastructure across the Emirate of Sharjah. Partnership aims to reinforce Sharjah's prominence in data management and sophisticated technologies while facilitating the emirate's transition toward a knowledge-centric economy.

Scale indicators6 records

Recent moves3 records

Expansion highlights6 records

Direct peers

• CoreWeave: Specialized GPU cloud provider offering NVIDIA-powered compute for AI training and inference. Highly comparable to DataCanvas as both focus on bare-metal and containerized GPU compute for AI workloads, with similar elastic scheduling and tiered GPU instance offerings.
• Lambda Labs: GPU cloud and deep learning infrastructure provider serving AI researchers and enterprises with on-demand GPU instances. Directly comparable product offering (cloud GPU clusters, multi-tier instances, model training and inference focus) to DataCanvas's Alaya NeW platform.
• Together AI: AI cloud platform providing GPU compute for training, fine-tuning, and inference of open-source large models. Comparable to DataCanvas in combining GPU-as-a-Service with model deployment tooling and Day0 support for frontier models like DeepSeek.
• RunPod: GPU cloud platform providing on-demand and serverless GPU instances for AI training and inference. Comparable self-serve model with container-based deployment and tiered GPU offerings, similar to DataCanvas's pay-per-use 'du' unit model.

Broad incumbents

• Alibaba Cloud (Aliyun): China's largest hyperscale cloud provider offering GPU compute instances (including H800), AI Platform for AI/ML workloads, and PAI for model training. Overlaps with DataCanvas on GPU compute and AI model development, though as part of a much broader cloud portfolio.
• Huawei Cloud: Major Chinese hyperscaler offering Ascend-powered AI compute, ModelArts platform, and Pangu models. Directly competes with DataCanvas for enterprise AI infrastructure budgets in China, especially under export-controlled GPU scenarios.
• Tencent Cloud: Chinese hyperscaler with GPU compute offerings and AI platform services. Comparable to DataCanvas as a competitor for Chinese enterprise AI workloads, with deeper capital resources and broader enterprise reach.

Emerging players

• Fireworks AI: AI inference platform offering optimized model serving for open-source LLMs including DeepSeek. Comparable to DataCanvas's large model inference platform and Day0 adaptation strategy for frontier model deployment.
• Modal Labs: Serverless AI compute platform for running AI workloads on GPUs with auto-scaling containers. Comparable to DataCanvas's elastic container cluster and serverless compute offerings for AI developers.
• Vast.ai: GPU marketplace connecting compute supply and demand across heterogeneous hardware. Conceptually aligned with DataCanvas's 'du' standardized computing unit vision for creating a tradable GPU compute marketplace.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds9 records

Investors21 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about DataCanvas