Choose the right chips, embodied intelligence robots, and computing power —Check real-world tests first, then order.
Bring parameter-based selection, independent reviews, and enterprise procurement together in one place. Make technical decisions with clear evidence.
In-Stock Marketplace
Product Selection & Procurement
Volcengine Ark model API (OpenAI SDK compatible)
D-Robotics RDK S100 Robotics Development Kit
Alibaba Cloud Bailian Non-Real-Time Speech Recognition (Qwen-Audio-ASR)
Large Models & Software ServicesVolcano Ark GUI Agent Capability
Reviews & Rankings
Public test methods and limitations
Industrial Thermal Imaging Module Temperature Difference Recognition Experience
Organizes test items and evaluation methods for temperature difference recognition of industrial thermal imaging modules.
Observation of Continuous Grasping Stability of a Parallel Gripper
An overview of test items and evaluation methods for the continuous grasping stability of a parallel gripper.
16-DOF Dexterous Hand Small Object Grasping Test
An overview of test items and evaluation criteria for small object grasping with the 16-DOF dexterous hand.
Review Rating Leaderboard
More ↗Sort by Published Report Score
01Compute and Procurement Requirements
Connecting compute suppliers with procurement demands
Compute Market
Compare configuration and delivery conditions in one table
| Specifications / Supplier | Region · Delivery | CNY / card-hour | Available cards | Request a Quote |
|---|---|---|---|---|
| NVIDIA A100 NVLink 40GB腾讯云 · 40 GB | Guangzhou, Shanghai, Nanjing, BeijingCloud server instance (GPU cloud server managed via CVM) | Pending Quote | Available Pending Confirmation | Request a Quote |
| NVIDIA V100尊云(广东尊云网络科技有限公司) · 32 GB | Nationwide 10 multiple availability zones, distributed across East China, North China, South China, Inner Mongolia and other regionsGPU Cloud Host (Cloud Instance, Computing Type) | Pending Quote | Available Pending Confirmation | Request a Quote |
| NVIDIA Tesla M60天翼云科技有限公司 · 8 GB | Tianyi Cloud availability zones equipped with NVIDIA M60 GPU cards (specific regions not listed on the page)Elastic Cloud Server (ECS) cloud server instance, GPU hardware passthrough, XEN virtualization type | Pending Quote | Available Pending Confirmation | Request a Quote |
| Nvidia 8卡A100-80GB商汤大装置 sensecore · 80 GB | Shanghai (cn-sh-01), Guangzhou (cn-gz-01), and other regionsElastic Bare Metal Server (BMS), physical cloud server, minute-level delivery, support for failover migration | Pending Quote | Available Pending Confirmation | Request a Quote |
Demand Hall
Find procurement and project cooperation opportunities
Open Tender for the Shandong College of Electronic Technology Integrated Circuit Teaching Large Model Project
公开招标 · Software · Jinan City
截止 12-07 09:30Open Tender for Hardware Procurement of the Fujian Sangang Steel Slag Transport Corridor Traffic Safety Control Project
公开招标 · Computing Power and Servers · Sanyuan District, Sanming City, Fujian Province
截止 11-24 15:00Public Tender for Hardware Equipment Procurement for the Public Digital Culture Platform of Baoding News Media Center
公开招标 · Computing Power and Servers · Baoding City, Hebei Province
截止 12-08 09:15Beijing JOINN Drug Assessment 2026 High-Performance Computing Server Procurement Open Tender
Public Tender
截止 12-04Insights & News
Technology Selection and Industry Insights
Yue Hongyuan A: Acquisition of Jinwangda Equity Still Pending Shareholders' Meeting Review on 11 30
Yue Hongyuan A (000573) replied on its investor interaction platform that the acquisition of Jinwangda equity is still pending review at the shareholders' meeting on 2026, 11 30. The company disclosed that Jinwangda specializes in precision transmission functional components, with core products including precision linear guide rails, ball screws, and planetary roller screws; its business composition is approximately 3C at 40%, new energy (lithium battery + PV) at 40%, display panels at approximately 6%, and others at approximately 14%.
He Xiaopeng Announces XPeng IRON Humanoid Robot to Be Released to Commercial Customers in Q2 Next Year
具身智能机器人Automakers Race into Robotics: GM Pushes In-Plant Automation, XPeng IRON Targets Mass Production and Store Deployment Next Year
具身智能机器人D-Robotics and Infinigence Announce Strategic Partnership to Drive Synergy Between Embodied Intelligence Open-Source Infrastructure and Chips
Zhaomin Technology Discloses Joint Development with Customers of Multiple Humanoid Robot Precision Components Moving Toward Gradual Mass Production
具身智能机器人Zongheng Co., Ltd. Outlines Low-Altitude AI Algorithm Catalog and Domestic Adaptation Progress at Earnings Briefing
具身智能机器人银河通用Galbot ET1获CES 2027最佳创新奖 为首个获奖双足人形机器人
Companies & Brands
Discover suppliers and partners
本末科技
On its official website, Direct Drive Tech (Benmo Technology) positions itself as a company specializing in robot power modules and complete robots, stating that it has built dual business lines around direct-drive technology: "core power modules + wheeled-legged robot systems." Its product lines include direct-drive power modules, robot joint modules, coreless (hollow-cup) motors, frameless torque motors, base motors, and fitness power modules, with solutions covering home appliances, industrial and commercial applications, fitness and entertainment, and embodied intelligence scenarios. Production and R&D are located in Songshan Lake, Dongguan; the website discloses two major production bases covering over 30000 square meters, equipped with 17 automated and semi-automated production lines, 3 flexible production lines, and one quality inspection center. For buyers, the company offers both core components and complete wheeled-legged robots, which can be considered for evaluating joint-drive replacements of traditional geared motor solutions. The website provides a selection tool and download center, but specific parameters, pricing, production capacity, and business registration information require separate verification.
上海星思半导体有限责任公司
Shanghai Starsmicro Semiconductor is a baseband chip company focused on 5G/6G communication technologies, providing full-scenario integrated space-terrestrial baseband chip platforms and solutions for terminal devices and smartphones, with products covering 5G/6G eMBB, RedCap, and NTN scenarios. According to the Taibo industry database, its industry applications target smartphone direct-to-satellite connectivity, satellite communication terminals, airborne communications, UAV ad hoc networks, eVTOL integrated sensing and communication, vehicular communication, smart cockpits, 5G FWA, emergency communications, trunked communications, industrial IoT, and other Internet of Things and satellite communication scenarios. Since its founding in 2020, the company has gone through multiple funding rounds: an angel round of 1 hundred million RMB in 2020 month 12 (Hillhouse Ventures), a Pre-A round of 4 hundred million RMB and a Pre-A+ round in 2021, a A round of 1 hundred million USD in 2 month 2022, two strategic financings in 2023, and a Series B round of 5 hundred million RMB in 2024 month 4, with investors including CEC-affiliated funds, CDH, Hillhouse Ventures, Matrix Partners, Walden International, and others. Suitable for procurement parties interested in satellite communication terminals, vehicular connectivity, and industrial IoT baseband solutions for further engagement; full business registration information and bidding records require login on the source platform and are not included on this page.
摩尔线程
Moore Threads is a full-featured GPU company whose publicly listed product lines cover integrated AI computing centers and computing servers (MTT KUAE, MTT SGX5000), full-featured GPU and graphics cards (MTT S5000, S4000, S3000, S2000, gaming GPUs MTT S80, S70), professional visual computing acceleration cards (MTT X300, S50), AI laptops and desktops (MTT AIBOOK, MTT AICUBE), and AI modules (MTT E300). On the foundational software side, it offers the MUSA SDK, drivers, Moore Perf System, and MUSACODE, along with AI training/inference suites, cloud-native GPU virtualization (vGPU), and scientific computing suites. Solutions target artificial intelligence (integrated AI computing centers, DigitalME digital humans, AI inference), embodied intelligence and edge computing, cloud and data centers (cloud PCs), high-performance rendering (digital twin and GIS, industrial design and manufacturing), and video acceleration (broadcasting and professional AV, video conferencing). It is suitable for procurement and deployment scenarios requiring domestic full-featured GPU computing power, AI computing clusters, or supporting software stacks; no specific pricing, qualifications, or customer lists are shown on the pages, so these should be verified separately before selection.
金山云
Kingsoft Cloud was founded in 2012 and listed on NASDAQ in the United States in 5 2020 (ticker: KC), describing itself as one of China's top three internet cloud service providers. Its publicly stated positioning is as a customer-centric cloud service provider, building its cloud computing infrastructure and operations on Kingsoft Group's enterprise-grade service experience. It combines technologies such as artificial intelligence, big data, IoT, blockchain, edge computing, and AR/VR with cloud computing, offering more than 120 solutions for industries including government, finance, healthcare, education, media, industrial, video, gaming, e-commerce retail, real estate, energy, and agriculture, serving 243 leading customers. Its data centers and operations are located in domestic regions such as Beijing, Shanghai, Guangzhou, Hangzhou, Yangzhou, and Tianjin, as well as international regions including the United States, Russia, and Singapore. For readers seeking suppliers, Kingsoft Cloud is suitable for cloud infrastructure for government and enterprise sectors, industry cloud solutions, and integrated technology scenarios involving AI/big data; this page is the company's official introduction, and solution details, pricing, and actual service capabilities need to be verified separately.
青云科技
QingCloud Technologies is a technology-leading enterprise cloud service provider and digital solutions vendor that launched public cloud services at its founding. According to its official blog QingCloud Chronicles, its public cloud has formed a strategic service portfolio centered on cloud platforms, cloud native, high-performance computing, big data, and audio/video. At the product level, offerings include e3 cloud servers equipped with 3rd Gen Intel Xeon Scalable processors, high-performance computing EHPC, real-time audio/video RTC, QKE container cloud service, and a big data workbench, with solutions for e-commerce, EHPC, audio/video, transportation and logistics, hybrid cloud, and other scenarios. In addition, QingCloud Marketplace was launched, aggregating nearly 300 products and services spanning basic software, website building, enterprise applications, artificial intelligence, and security applications. This is a useful reference for enterprises seeking suppliers of cloud platform, cloud native, high-performance computing, and industry cloud solutions; this article is vendor self-reported content, and product performance and coverage should be understood per the vendor's own claims.
曙光信息产业股份有限公司
Sugon Information Industry Co., Ltd. (referred to on its official website as "Sugon", SSE code 603019) is publicly positioned as a computing power and cloud services company serving the digital transformation of governments and enterprises. Its business covers servers (rack-mounted, high-density, and core-application servers), high-performance computers, data center cooling and liquid-cooling infrastructure, storage products, cybersecurity products, intelligent acceleration cards, edge computing platforms, big data platform software, and cloud computing services (ECS cloud servers, BMS bare metal servers, OSS object storage, CCI cloud container instances, etc.). Its cloud computing product line includes Sugon Cloud HCI hyper-converged software, Sugon Cloud Stack full-stack cloud platform, and Sugon Cloud SVM virtualization software, and it promotes a "domestic chip-level SoC root of trust" and full-stack trusted computing capabilities. Its industry solutions target scenarios such as the internet, telecom operators, electric power, petroleum, scientific research, education, finance, transportation, and healthcare. The information is mainly sourced from the company's official website homepage and product navigation pages and does not include verification details such as business registration or founding date.
猎户星空(OrionStar)
OrionStar (猎户星空) describes itself on its official website news page as founded in 2016 in 9, positioning itself as an AI robot service-oriented solution provider. It proposes the formula "AI + Software + Hardware + Services = Robot" and claims to possess proprietary full-stack AI technology, robot operating system application development capabilities, standardized hardware R&D and manufacturing, and cloud-based brain services. Its publicly disclosed business lines include the Robot business and Smart Mobility business. In 2023, the company announced a strategic upgrade: alongside AI robots as its core business, it positioned "Juyan," built on large model technology, as its second growth curve, offering digital employees to enterprise clients to improve operational efficiency. For readers seeking suppliers, relevant cooperation scenarios include deploying voice interaction robots (such as the Leopard Xiaomi series) for guided tours, reception and greeting, and wayfinding guidance, as well as large model digital employee services. The official website also lists a consultation phone number and a cooperation email as contact channels. It should be noted that the above information comes from the company's own official website promotional pages; service capabilities and customer scale figures have not been verified by third parties.
上海燧原科技股份有限公司
Shanghai Enflame Technology Co., Ltd. (Enflame Technology) is positioned as an AI neural network solutions provider, focusing on cloud computing power platforms in the artificial intelligence field. Its core products are high-end deep learning chips designed for cloud data centers, targeting AI training platforms. According to the Taibo Network industry database, its chips adopt a self-developed architecture, featuring high computing power, high energy efficiency, programmability, and low cost, while supporting mainstream machine learning frameworks, and are designed and optimized for cloud AI training. The company aims to provide domestic general-purpose AI software and hardware solutions for cloud AI training and inference, supplying computing power to cloud service providers including public, private, and hybrid clouds, and serving industries such as finance, healthcare, and education as well as intelligent data centers under "New Infrastructure." Funding information shows it has reached Series E and beyond; 2025-05-14 investors include SDIC Juyi, Shanghai Guotou (Shanghai State-Owned Capital Investment), and Shanghai International Group. This page is an industry database entry; product specifications and delivery capabilities should be further verified against the company's official website before procurement and selection.
智平方(深圳)科技股份有限公司
In its exhibitor profile for the 2026 World Robot Conference, Zhi Ping Fang (Shenzhen) Technology Co., Ltd. (智平方) publicly positions itself as an AGI-native productivity-focused general-purpose humanoid robotics company, deeply engaged in the high-barrier, AI-intensive embodied intelligence space, building an integrated capability system of "model + hardware + scenarios." Its core technology is the fully in-house developed embodied foundation model AlphaBrain, built with native AI advantages for the real physical world, with a series of end-to-end VLA foundation models and related full-stack capabilities open-sourced. Its flagship product, the AlphaBot series, is positioned as a mass-production-grade general-purpose humanoid robot; the exhibitor introduction states it has been deployed in high-value scenarios such as automotive manufacturing, semiconductors, and public commercial services, with orders from benchmark customers. The exhibit AlphaBot 2 is described as a robot form factor defined by the AlphaBrain embodied foundation model, emphasizing spatial intelligence and the ability to perform tasks in novel environments; the application case page states that it has reached a strategic partnership with an automaker, with AlphaBot 2 to perform tasks in quality inspection, assembly operations, logistics transfer, and other production processes. From a procurement and partnership perspective, the company's public selling points center on a closed-loop narrative of "foundation model + complete unit + scenario data feedback," making it suitable for buyers focused on embodied intelligence deployment in industrial scenarios such as automotive manufacturing and semiconductors to further verify its actual delivery and deployment scale. Business registration details such as registered capital, founding date, and credit code, as well as specific customer names, are not provided in the source and require separate verification.
润泽科技发展有限公司
Runze Technology Development Co., Ltd. (referred to as Runze Technology) is a member of the Data Center Branch of the China Computer Users Association, positioned as a comprehensive service provider for the design, construction, and operation of intensive, large-scale big data center industrial clusters. The company adopts a self-invested, self-built, self-owned, and self-operated model, with business covering the investment, R&D, construction, and operation of intelligent computing power and computing infrastructure. According to the association's page, the company was listed in 2022 8, with stock code 300442, and claims to have deployed AIDC (AI data center) clusters across seven major nodes and eight campuses nationwide, primarily serving digital economy industries such as the internet, big data, cloud computing, artificial intelligence, and blockchain. For readers seeking suppliers, its applicable scenarios are construction and O&M cooperation for hyperscale computing infrastructure and AIDC machine room clusters. It should be noted that this page is an association member profile and does not provide verifiable specific technical parameters such as machine room scale, rack counts, or power capacity.
北京九章云极科技股份有限公司
DataCanvas (Jiuzhang Cloud Extreme) positions itself as a provider of AI infrastructure and intelligent computing cloud services. Guided by an "AI Factory" strategy, its core businesses include Jiuzhang Intelligent Computing Cloud (Alaya NeW Cloud), Jiuzhang Intelligent Computing OS (Alaya NeW OS), computing power package services, AI Factory, the Alaya Token delivery platform, and a Developer Center for developers. The company provides computing power scheduling and intelligent computing cloud services for AI training and inference, was the first to propose the DCU (one degree of computing power) measurement standard, and operates a distributed intelligent computing cloud with tens of P-scale (petaFLOPS-level) intelligent computing capacity, forming a "central scheduling + regional coordination" architecture. Its intelligent computing cloud platform offers model training, inference, and management services, with computing power packages delivered via a Serverless architecture in three series: A/E/B. Its open-source ecosystem includes the DingoDB multimodal vector database, YLearn causal learning software, and DAT automated machine learning software. It is suitable for buyers seeking intelligent computing cloud, computing power packages, and MaaS-type services; specific qualifications and customer partnerships are subject to official disclosures.
爱芯元智半导体股份有限公司
Axera Semiconductor Co., Ltd. is a high-tech company headquartered in Ningbo, founded in 2019, specializing in the research and development of high-performance, low-power AI edge-side inference chips. The company adopts a self-developed core IP strategy, with core technologies including Axera AI-ISP and Axera Mixed-Precision NPU, providing key computing power and perception capabilities for scenarios such as assisted driving, terminal computing, and edge computing. According to reports on the company's official website, it has achieved the development and mass production of multiple generations of chips since its founding, and in 12 2025, was selected for Zhejiang Province's "Sci-Tech Rising Star" company list (selected from nearly 400 candidate companies, with selection requirements of an annual revenue or valuation growth rate of over 30% in the past three years). Suitable for companies seeking edge-side AI computing chips, intelligent driving, and edge computing solutions to connect with its automotive or general-purpose computing business lines. This introduction is based solely on the official website news page; business registration information, customer lists, and product test data have not been verified.
深圳乐动机器人股份有限公司
Shenzhen LDROBOT Co., Ltd. focuses on the independent R&D and manufacturing of robot mobile perception technology, publicly positioning itself as a supplier of sensors and solutions for mobile robots. Its technology system is built around three core services: a full sensor lineup, proprietary algorithms, and a cloud platform, with product lines covering DTOF LiDAR, triangulation LiDAR, solid-state line-scan LiDAR, single-point LiDAR, and ultrasonic sensors. The source states that average annual shipments of its triangulation LiDAR exceed the million-unit level, and that it has launched industry-first products such as the consumer-grade MINI DTOF LiDAR. Its application scenarios center on the perception stage of mobile robots, including home service robots, commercial service robots, smart lawn-mowing robots, drones, logistics and warehouse robots, and ROS education. At this CIOE China International Optoelectronic Exposition Intelligent Sensing Expo, the company is exhibiting at Booth 6C21 in Hall 6, with exhibits of robot perception sensors; its listed application fields are consumer electronics/entertainment (home appliances) and sensing and test & measurement (smart sensors). Fit points for buyers: when selecting perception sensors (especially LiDAR solutions) for mobile robots or home appliances, requirements can be matched against the product lines above; the source is an exhibitor page for the expo and does not provide specific specifications, certification numbers, or contact details, so product details should be verified with the company before selection.
沐曦集成电路(上海)股份有限公司
MetaX Integrated Circuits (Shanghai) Co., Ltd. (stock code: 688802.SH) was founded in 2020-9 and listed on the STAR Market in 2025-12, with headquarters in Shanghai. The company focuses on self-developed full-stack high-performance GPU chips and computing platforms, providing computing power for AI training, inference, graphics rendering, AI for Science, and other scenarios. Its publicly disclosed product lines cover four major GPU series: the Xisi N Series (inference), the Xiyun C Series (training and general-purpose computing, including C500, C550, C500X, C588, C600), the Xicai G Series (graphics rendering), and the Xisuo X Series (the inference line is categorized according to website navigation), complemented by the self-developed MXMACA heterogeneous computing software stack, emphasizing software-hardware co-design. At the system level, the company offers GPU servers, GPU workstations, and GPU supernodes (such as the Xiyun C550 Shanghai Cube liquid-cooled rack and the Xiyun C500X optical-link supernode, as well as the Xiyun C550 3D Mesh supernode), and its self-developed MetaXLink high-speed interconnect technology supports scaling up to 10,000-GPU clusters. Industry solutions span finance, healthcare, energy, education and research, transportation, and digital entertainment, with additional布局 in emerging fields such as aerospace computing power, embodied intelligence, and the low-altitude economy. For procurement decision-makers, vendor selection can proceed along the chain of "accelerator cards — servers/supernodes — MXMACA software ecosystem" when reviewing public materials, while noting that its ecosystem differs from the CUDA ecosystem and migration costs need to be assessed.
北京芯驰半导体科技股份有限公司
SemiDrive provides automotive-grade SoCs and high-performance MCUs for next-generation automotive E/E architectures, covering application scenarios such as smart cockpit, smart gateway, and core vehicle control; the company is also actively expanding into embodied intelligence and industrial control, forming a dual-driven business layout spanning automotive and robotics. Its public product lines include cockpit processors (X9 series and X10 AI cockpit processors), gateway processors (G9X/G9Q/G9H), vehicle control MCUs (E3 series), and industrial and embodied intelligence products (industrial applications, embodied intelligence "brain" R1 series, "cerebellum" D9 series, motion control hub E36 series, joint modules and dexterous hand E3118-R, and LiDAR E3118). The company has R&D centers in Beijing, Shanghai, Nanjing, and Shenzhen. For buyers seeking automotive-grade chip or embodied intelligence chip suppliers, the company offers a complete portfolio of chips and modules ranging from cockpit, gateway, and vehicle control MCUs to robot brain/cerebellum/joint modules, making it suitable for sourcing comparisons across both automotive electronics and complete robot scenarios; the footer lists the company name as "Beijing SemiDrive Semiconductor Technology Co., Ltd."
芯海科技(深圳)股份有限公司
ChipSea Microelectronics was founded in 2003/9. It is a full signal-chain IC design company built on a dual technology platform of "Analog + MCU", with business coverage spanning sensing, computing, control, power, connectivity, and AI technology platforms. Its chip and algorithm solutions serve smart health, communications and computing, lithium battery management, automotive electronics, industrial measurement, premium consumer electronics, and AI edge intelligence hardware. Key considerations for buyers: the company claims to be an EC chip supplier certified on both Intel and AMD platforms in the computing ecosystem segment; in automotive electronics it offers ASIL-D automotive-grade MCUs and SmartAnalog industrial-grade signal conditioning products; and in battery management it has a BMS product line targeting the premium consumer market. Headquartered in Shenzhen, the company has subsidiaries in Beijing, Shanghai, Chengdu, Hefei, Xi'an, Hong Kong, and Singapore. Note: all positioning and industry-standing statements above are self-reported by the company on the source page. This page is a brand-operated official store on the Huaqiang Electronics platform; the relevant certifications and rankings have not been independently verified on this site.
矽品科技(苏州)有限公司
According to business registration information on Qichamao, Siliconware Technology (Suzhou) Co., Ltd. was established on 12 2001, 29. It is a wholly foreign-owned limited liability company, currently in operation (active status), located at No. 288 Fengli Street, Suzhou Industrial Park. Its registered business scope covers the assembly of micro camera modules and LCD display modules; the R&D, design, manufacturing, packaging, testing, processing, and repair of memory modules, flash memory cards, semiconductor integrated circuits, transistors, electronic components, electronic materials, and related products; as well as the sale, wholesale, commission agency, import/export, and supporting services of the products it produces; in addition, general projects including leasing of non-residential real estate and machinery and equipment. Based on the registration information, this company is well suited for manufacturing and supporting partnerships in semiconductor packaging and testing, module assembly, and processing. Please note that the data on this page comes from a third-party enterprise information platform; the company's business capabilities, production capacity, and qualifications still need to be verified via its official website or other first-hand sources.
埃夫特智能机器人股份有限公司
Eft Intelligent Robot Co., Ltd. is a provider of industrial robot complete machines and solutions. Public sources indicate that its business covers processes such as spraying, welding, palletizing, and material handling, and it offers specialized solutions for electronics manufacturing, including automatic screw-driving workstations, intelligent flying-photography workstations, high-cleanliness SCARA robots, and positive-pressure explosion-proof robots. Its subsidiary Qizhi (Wuhu) Intelligent Robot Co., Ltd. focuses on embodied intelligence, building a general-purpose technology foundation for intelligent robots composed of Modou IDE, the Dayan data platform, and OpenmindOS, along with the yobot series of humanoid and wheeled robot platforms, and has demonstrated multi-robot collaborative operations in retail shelf-stocking scenarios. In 2026, its industrial robot products obtained functional safety certification from SGS based on ISO 13849 PL d, Cat.3, and its large-payload robot ER300-2700 was selected as one of the second batch of Anhui Province's First (Set) Major Technical Equipment items of 2025. The company operates multiple subsidiaries in Europe, including CMA ROBOTICS and ROBOX, and has established localized sales and after-sales networks in markets such as Italy, Germany, and France. Suitable for reference by parties seeking partnerships in industrial robot complete machines, large-payload robots, and embodied intelligence platforms; this page is a company-owned news page, and business registration information and third-party evaluation data need to be verified separately.
珞石(山东)机器人集团股份有限公司
ROKAE Robotics was founded in 2015 and is publicly positioned as a "next-generation intelligent robotics expert", focusing on integrated "Robot + AI" innovation. Its product line covers multi-joint industrial robots, flexible collaborative robots, and humanoid robots, among other embodied intelligence products, providing automation solutions for industrial, commercial, and medical sectors. The company is headquartered in Shandong, with robotics R&D centers in Beijing, Tokyo, Shanghai, and Shenzhen (the English introduction also mentions Wuhan). It is participating in the Shenzhen (International) General AI Industry Expo as an exhibitor, at booth 12D106. For readers looking for robotics suppliers, product specifications and capacity data can be further verified at the expo or on the official website; this page is an expo exhibitor information page, and qualifications and customer information have not been independently verified.
壁仞科技
The Research Institute page on Biren Technology's official website shows that the company conducts research around AI computing power needs and chip enabling technologies, with cutting-edge algorithm research, advanced architecture innovation, and full-stack collaborative optimization as its main directions. Public activities involve the GPU product ecosystem, including collaborating with Shanghai Jiao Tong University to explore AI-Native GPU code generation, integrating into the DeepSpeed ecosystem to enable cross-hardware migration of large model training and inference, and co-hosting MeetUps with the OpenAnolis (Longxi) community and the PyTorch ecosystem. Based on available sources, the company is positioned as a vendor of AI computing power chips and supporting software stacks, with its Research Institute focused on algorithms and hardware-software co-optimization. Limitations of the information: current sources are limited to the Research Institute's papers/updates page and do not include the company's product models, specifications, qualifications, or business information.
云知声
According to the "About Unisound" page on the company's official website, Unisound was founded in 2012 and describes itself as a China-native large model company with multimodal large models as its core foundation, industry agents as its key lever, and large-scale deployment in serious (mission-critical) scenarios as its distinguishing feature. It listed on the Hong Kong Stock Exchange in 2025 (stock code: 09678.HK). Its website states it has formed a technology closed loop of "self-developed chips + proprietary AI computing + native models + serious scenarios." Its public product lines and business directions include: the Shanhai large speech model (launched in 2023) and the native agent large model U2 (released in 2026); consumer-grade AI chips (over 600 units in 2021, approximately 1300 units in 2022) and the UniOne series of edge conversational AI chips; NPU IP licensing for automotive chips; medical solutions for hospitals (the website says deployed in over 200 hospitals nationwide); metro smart living solutions covering cities such as Guangzhou, Shenzhen, Qingdao, and Nanning; as well as healthcare IT products such as a medical insurance supervision platform and medical record quality management. Headquartered at Jinyu Intelligent Manufacturing Park, No. 27 Jiancaicheng Middle Road, Haidian District, Beijing, the company also has subsidiaries or offices in Shanghai, Shenzhen, Hangzhou, Xiamen, Hong Kong, and other locations. In terms of collaboration scenarios, the above deployment directions suggest it is suitable for algorithm and chip solution partnerships in scenarios such as smart healthcare, smart transportation, smart cockpit, and consumer smart hardware. Limitations: this page is a company self-description and contains no third-party verification or specific product pricing.
Roborock 石頭科技(台灣官網)
Roborock's official Taiwan website laundry category page showcases its washer-dryer product line, including two flagship smart washer-dryer combo models, the Roborock H1 and H1 Neo, both featuring Zeo-cycle® molecular sieve low-temperature drying technology and a self-cleaning system, with support for smart control via the Roborock APP; there is also the compact Roborock M1 Pure, positioned for all-around multi-scenario care. For readers, this page shows that beyond its robot vacuum business, Roborock offers washer-dryer combo products with self-cleaning and app connectivity in the laundry appliance segment, making it worthwhile for buyers interested in premium washing-drying and low-temperature drying technology to further compare the configuration differences between the H1 and H1 Neo (specific differences should be checked on each product detail page). This page does not provide pricing, after-sales service, or qualification information, which should be verified separately before purchasing.
云从科技集团股份有限公司
CloudWalk Technology Co., Ltd. presents on its official website a product system spanning three main lines: perception devices, platform systems, and large models. On the perception side, offerings include human-machine collaboration devices such as the Dige AI-Defined Mini Station and the Congrong Large Model Training & Inference All-in-One Machine, as well as AIoT devices including panel machines, video face access-control machines, desktop ID verification devices, the Polaris structured light module, the Proxima liveness security module, and multiple smart zoom cameras. On the platform side, offerings include the Bailan Data Governance Platform, the Kaiming Privacy Computing Platform, the human-machine collaboration operating system CWOS, the Xingyun Agent Operating System, the Galaxy AI Platform, a smart IoT platform, and an intelligent customer service platform, along with large-model-related products such as Mofang Agent Lab, Yunqi Model Hub, the Congrong Large Model, and DataGPT. Business areas cover industry development directions such as smart finance, smart energy, and smart manufacturing, as well as Digital China directions including smart governance, smart mobility, smart customs, smart airports, smart transportation, smart cities, and grassroots governance. In the AI agent space, the company provides capabilities such as human-machine dialogue, document assistant, knowledge engine, intelligent image creation, and digital humans. The official website provides a business cooperation email and a 400 hotline.
云从科技集团股份有限公司
CloudWalk Technology Group Co., Ltd. is a company publicly positioned as a "provider of AI infrastructure and AI team solutions." It was incubated by the Chinese Academy of Sciences in 2015 and trades under the stock code 688327. Its business is built around "AI infrastructure + AI teams": on one hand, it provides model and computing power infrastructure through the Congrong multimodal large model, training-inference all-in-one machines, trusted data spaces, and more; on the other hand, through the Xingyun Agent Operating System, Cube Agent Lab, and Yunqi Model Hub, it organizes models, data, and tools into agent teams for business functions such as content production, marketing, enterprise operations, and R&D. Its public product lines also include intelligent AIoT devices such as facial recognition terminals, structured light liveness modules, and smart cameras, as well as platform systems for privacy computing, data governance, and smart IoT, applied across industries including finance, governance, transportation, energy, and manufacturing. For buyers seeking large model platforms, agent systems, or AIoT sensing devices, the company offers multi-layered supply from edge devices to training-inference platforms; specific product parameters, pricing, and delivery capabilities should be verified separately.
本末科技◇上海星思半导体有限责任公司上海星思半导体有限责任公司◇摩尔线程摩尔线程◇金山云金山云
青云科技◇曙光信息产业股份有限公司曙光信息产业股份有限公司◇猎户星空(OrionStar)猎户星空(OrionStar)◇上海燧原科技股份有限公司上海燧原科技股份有限公司◇智平方(深圳)科技股份有限公司智平方(深圳)科技股份有限公司◇润泽科技发展有限公司润泽科技发展有限公司◇北京九章云极科技股份有限公司北京九章云极科技股份有限公司◇爱芯元智半导体股份有限公司爱芯元智半导体股份有限公司◇深圳乐动机器人股份有限公司深圳乐动机器人股份有限公司◇沐曦集成电路(上海)股份有限公司沐曦集成电路(上海)股份有限公司◇北京芯驰半导体科技股份有限公司北京芯驰半导体科技股份有限公司◇芯海科技(深圳)股份有限公司芯海科技(深圳)股份有限公司◇矽品科技(苏州)有限公司矽品科技(苏州)有限公司◇埃夫特智能机器人股份有限公司埃夫特智能机器人股份有限公司◇珞石(山东)机器人集团股份有限公司珞石(山东)机器人集团股份有限公司◇壁仞科技壁仞科技◇云知声云知声
Roborock 石頭科技(台灣官網)◇云从科技集团股份有限公司云从科技集团股份有限公司◇云从科技集团股份有限公司云从科技集团股份有限公司品牌入驻展示在这里虚位以待Procurement & Partnerships
Membership, Merchant & Brand Services
Enterprise Procurement
Manage product selection, BOM items, inquiry and quotation progress.
Enter Member Center ↗Supplier ServicesMerchant Onboarding
Display products by category, respond to procurement requests, and manage supply documentation.
Apply for Partnership ↗Brand ServicesBrand Review Program
Disclose sample sources and test methods, and provide verifiable measured data.
Learn about sample testing ↗