What Aivres does

Aivres is a data center and AI infrastructure solutions provider founded in 2018 and headquartered in Milpitas, California (Silicon Valley), operating as a U.S. subsidiary of Chinese systems vendor Inspur. The company designs and manufactures a full-stack AI infrastructure portfolio spanning rack-scale AI supercomputers based on NVIDIA GB200/GB300 NVL72 and Vera Rubin NVL72 reference architectures (KRS8000, KRS8500V3, KRS8000V4), 5U-6U HGX-class AI servers (KR6288, KR9288, KR5288, KR6268, KR6278), general compute and multi-node servers, high-density storage servers, networking (InfiniBand and Ethernet), and an end-to-end liquid cooling portfolio that includes in-rack CDUs at 100-240kW and in-row 1.3MW units achieving PUE as low as 1.15. The portfolio is delivered alongside two software platforms, KSManage for infrastructure orchestration and AIStation for AI resource scheduling.

Aivres makes money primarily through hardware sales of purpose-built servers and liquid cooling systems, supplemented by Joint Design Manufacturing (JDM) professional services that co-design customized platforms with customers. The company goes to market through direct enterprise field sales, a Diamond-sponsorship presence at NVIDIA GTC and OCP Global Summit, regional demo centers in the US and Southeast Asia, and NVIDIA Technical Workshops in Malaysia and Paris. Its stated customer base includes six of the world's top ten cloud service providers, with named deployments at Baidu (10,000+ nodes in 8 hours) and ByteDance / Aolani Cloud (approximately 36,000 NVIDIA B200 chips in Malaysian data centers, valued at around US$2.5 billion). The company operates across 113 countries with 8 global R&D centers and 6 global production centers, and its JDM model compresses new product development cycles from an industry-standard 1.5 years to 9 months.

The company has flagged as a U.S. subsidiary of Inspur, which creates a notable geopolitical dimension to its commercial profile, particularly given the ByteDance / Aolani Cloud engagement that supports Chinese technology companies in deploying AI infrastructure in overseas locations to navigate US export controls. Key technology partnerships span NVIDIA (flagship), AMD and Intel (core processor partners), and VAST Data (strategic software-defined storage integration announced March 2026). Pricing is quote-based and not publicly disclosed, consistent with a large-enterprise B2B hardware sales model.

Name

Aivres

Legal name

Aivres

Website

https://aivres.com

Company type

Private

Founded year

2018

Operating status

Operating

Headcount range

251–500 employees

Short description

Aivres is a U.S.-headquartered AI infrastructure provider (Inspur subsidiary) that designs and manufactures GPU servers, rack-scale AI supercomputers, and liquid cooling systems for hyperscale data centers, cloud service providers, and enterprise AI customers globally.

Ownership category

akta.pro rank

Product category

AI Data Center Server Hardware

NAICS

Electronic Computer Manufacturing (334111), Computer and Peripheral Equipment Manufacturing (33411), Computer Systems Design and Related Services (54151)

SIC

Electronic Computers (3571), Services-Computer Integrated Systems Design (7373)

akta.pro primary industry

AI Server Systems & HGX/Accelerator Platforms (HDAAAAAB)

akta.pro secondary industries

AI Datacenter Power, Cooling & Racks (liquid cooling, power delivery) (HDAAAAAH), AI Integration & Orchestration Platforms (Connectors, Workflow, iPaaS for AI) (HDAEANAI)

Keywords

• AI GPU servers
• Data center infrastructure
• Liquid cooling solutions
• Rack-scale computing
• Cloud server hardware

Headquarters

HQ city

San Jose

HQ country

United States

HQ region

North America

Offices3 records

Markets served

GTM type

B2B

Offering type

Hardware or Manufacturing

Cost components

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

Revenue model

1. Hardware Server Sales: Sale of purpose-built rackmount servers including compute servers, AI/GPU servers, multi-node servers, and storage servers. Products are sold directly to hyperscale data centers, cloud service providers, and enterprise customers.
2. Liquid Cooling Solutions: CDU (Coolant Distribution Unit) products and liquid-cooled rack solutions sold alongside servers or as standalone cooling infrastructure for AI data centers.
3. Joint Design Manufacturing Services: Custom product development and manufacturing services through JDM model, enabling close collaboration with customers on customized hardware solutions from design through deployment.

Go-to-market motion2 records

Distribution channels3 records

Marketing channels7 records

Core offering

Aivres designs, manufactures, and deploys data center and AI infrastructure hardware — including rack-scale AI supercomputers, AI/GPU servers, compute servers, multi-node servers, storage servers, and end-to-end liquid cooling solutions — sold directly to hyperscale data centers, cloud service providers, and enterprise AI customers across 113 countries. The company complements this hardware with proprietary infrastructure management and AI resource orchestration software (KSManage, AIStation) and a Joint Design Manufacturing (JDM) model that enables customized, rapid deployment of purpose-built AI factories.

Product overview

Aivres offers a comprehensive portfolio of data center and AI infrastructure solutions consisting of AI/GPU servers (KRS rack-scale systems, KR AI servers with NVIDIA HGX platforms, RTX PRO servers), general compute servers (KR1280, KR2280), multi-node servers (K24 series), and supporting infrastructure (liquid cooling solutions, open infrastructure, networking). The portfolio spans from enterprise single-server deployments to hyperscale rack-scale AI supercomputers, with software platforms including KSManage for infrastructure management and AIStation for AI resource orchestration. Products support NVIDIA, AMD, and Intel processor/GPU technologies with various cooling options (air, liquid, hybrid) to address diverse data center requirements.

Differentiator

Problem solved

Functional benefit

Products and services

• KRS8000V4 Rack-Scale AI Supercomputer Rack-scale AI supercomputer supporting NVIDIA Vera Rubin NVL72, built for trillion-parameter LLM training and emerging agentic AI workloads in hyperscale data centers.
• KRS8500V3 Liquid-Cooled Exascale AI Rack Liquid-cooled exascale AI rack built on the NVIDIA Blackwell Ultra architecture, designed for extreme AI compute workloads at hyperscale.
• KRS8000 Liquid-Cooled AI Rack Liquid-cooled AI rack for trillion-parameter LLM training based on the NVIDIA GB200 NVL72 platform, delivering 30x faster real-time LLM inference and 25x the performance of H100 air-cooled infrastructure.
• KR6288 6U AI Server 6U AI data center server with NVIDIA HGX H200 8-GPUs delivering 32 PFlops of AI performance for hyperscale AI training and inference workloads.
• KR9288 10U AI Server 10U server with NVIDIA HGX B300 8-GPU and powerful air cooling for Blackwell Ultra AI workloads in data centers that prefer air-cooled deployments.
• KR5288 5U Hybrid Air/Liquid-Cooled AI Server 5U flexible hybrid air/liquid-cooled platform with NVIDIA HGX B300 8-GPU for high performance density in data centers requiring flexibility between air and liquid cooling.
• KR5288 Rubin AI Server 5U flexible hybrid air/liquid-cooled platform for traditional data centers supporting NVIDIA HGX Rubin NVL8 GPUs.
• KR2288 2U Ultra-Dense Liquid-Cooling AI System 2U ultra-dense liquid-cooling AI system for next-generation AI workloads, supporting NVIDIA HGX Rubin NVL8 GPUs in a compact form factor.
• KR6268 6U Enterprise AI Server 6U AI server featuring 8 NVIDIA RTX PRO 6000 Blackwell GPUs, delivering robust compute power, memory, and throughput enterprises require for the AI industrial revolution, including LLM training and advanced AI graphics.
• KR6278 6U Server 6U server supporting NVIDIA RTX PRO 4500/6000 Blackwell Server Edition GPUs, targeted at enterprise AI workloads and graphics-intensive applications.
• KR1280 1U General Compute Server Versatile flagship 1U 2-socket general compute server supporting Intel Xeon 6 or AMD EPYC 9005 processors, designed for high-density performance in mainstream data center workloads.
• KR2280 2U General Compute Server Versatile flagship 2U 2-socket server with exceptional compute performance, storage scalability, and flexible front/rear configurations for diverse enterprise workloads.
• K24-2S 2U 4-Node Multi-Node Server 2U 4-node 2-socket high-density server with liquid cooling for HPC and enterprise workloads, maximizing compute density in a compact footprint.
• KR1180V2 High-Density Storage Server High-density storage server integrated with VAST Data AI OS to accelerate AI-driven multi-exabyte storage deployments for hyperscale and enterprise AI customers.
• Liquid Cooling Solutions End-to-end liquid cooling portfolio including in-rack CDU products (100–240kW), in-row megawatt-class CDUs (1.3MW), cold-plate, EVAC, and immersion cooling — achieving PUE as low as 1.15 with over 80% cold plate coverage.
• Networking Solutions (InfiniBand and Ethernet Switches) High-speed data center networking portfolio including InfiniBand and Ethernet switches to interconnect AI and compute infrastructure.
• KSManage Infrastructure Management Platform Proprietary infrastructure management software platform for data center resource orchestration, monitoring, and management across compute, storage, and networking resources.
• AIStation AI Resource Platform Proprietary AI resource platform for managing training and inference workloads, increasing AI compute resource utilization from 40% to 80% in enterprise environments.
• NVIDIA-Powered AI Solutions AI platforms and server solutions featuring NVIDIA GPU and accelerator technologies across HGX H200, Blackwell, Blackwell Ultra, Rubin, Vera Rubin, and RTX PRO families.
• AMD-Powered AI Solutions AI servers and platforms powered by AMD Instinct GPUs (including MI350 series) and AMD EPYC 9005 processors for AI and HPC workloads.
• Open Infrastructure Solutions Open compute solutions following OCP standards for cloud, storage, enterprise and edge computing deployments.

Quantifiable outcome

• 30x faster real-time LLM inference and 25x performance of NVIDIA H100 at equivalent power
• +4 more outcomes

Named customers4 records

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration1 record

AI capability4 records

Feature7 records

Partnerships

Six partnerships are on record, tiered strategic, flagship, core and minor.

• VAST DatastrategicTechnology or Integration · 18 March 2026Aivres announced partnership with VAST Data to integrate VAST AI OS with Aivres KR1180V2 high-density storage server. The integrated solution aims at accelerating AI-driven multi-exabyte storage deployments by combining VAST's advanced software architecture with Aivres' hardware engineering and localized deployment capabilities.
• NVIDIAflagshipTechnology or IntegrationNVIDIA is Aivres' primary technology partner for GPU solutions. Aivres is a Diamond sponsor at NVIDIA GTC events and showcases AI Factory solutions powered by NVIDIA technology. Aivres develops rack-scale liquid cooling solutions based on NVIDIA GB300 NVL72 and Blackwell Ultra GPUs, including KRS8000, KRS8500V3, KR9288, and KR5288 products. Aivres' KR6268 supports NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs.
• AMDcoreTechnology or IntegrationAMD is a key technology partner for GPU and processor solutions. Aivres offers AI servers supporting AMD Instinct MI350 series GPUs and AMD EPYC processors across the server portfolio including KR6288-E3 with AMD EPYC 9005.
• IntelcoreTechnology or IntegrationIntel is a technology partner providing Xeon processors. Aivres servers support Intel Xeon Scalable and Intel Xeon 6 processors across the product line including V2 and V3 series.
• Aolani CloudstrategicStrategic or Co-development PartnerSoutheast Asian cloud provider partnering with ByteDance and Aivres for AI infrastructure deployment. Aivres is assembling approximately 500 Nvidia Blackwell computing systems with 36,000 B200 AI chips in Malaysian data centers.
• Bain Capital (Bridge Data Centers)minorChannel Partner/ Reseller/ DistributorBain Capital's Bridge Data Centers terminated contract with Megaspeed, reallocating 68.4 MW of power capacity. Megaspeed had acquired Nvidia H100 and H800 chips through Aivres Systems, a U.S. subsidiary of Chinese company Inspur.

Scale indicators12 records

Recent moves6 records

Expansion highlights5 records

Direct peers

• Supermicro: Closest direct peer: a leading provider of GPU-optimized AI servers (HGX H100/H200/Blackwell, liquid-cooled rack-scale systems) sold to hyperscalers and enterprises, with a similar NVIDIA-centric go-to-market.
• Inspur: Aivres' parent company and one of the world's largest server ODMs; competes directly in AI rack-scale systems, especially in China and EMEA, and shares manufacturing and supply chain with Aivres.
• Quanta Computer: Taiwan-based ODM building HGX-based AI servers for hyperscalers (including direct white-box shipments); a key direct competitor in the same JDM/ODM tier as Aivres.
• Wiwynn: Wistron subsidiary specializing in hyperscale and cloud AI server ODMs (including liquid-cooled NVIDIA platforms); a direct peer competing for the same CSP customers.
• Foxconn (Hon Hai Precision): Largest global ODM, expanding aggressively into NVIDIA HGX/Blackwhite AI rack-scale systems for hyperscalers; competes with Aivres on mega-deal scale and global manufacturing.
• GIGABYTE Technology: Provides G-series GPU servers (G593, G894) and rack-scale AI platforms for hyperscalers and enterprises; competes with Aivres in the NVIDIA HGX/Blackwell tier.
• ASUS: Offers ESC AI servers (ESC N8A, ESC8000 series) targeting hyperscale and enterprise AI workloads, with growing rack-scale offerings; a direct peer at the NVIDIA-based AI server tier.

Broad incumbents

• Dell Technologies: Sells PowerEdge XE9680/XE8640 AI servers and PowerScale storage to hyperscalers and enterprises; a broad incumbent that competes with Aivres on rack-scale AI deployments but across a much wider portfolio.
• Hewlett Packard Enterprise (HPE): Offers ProLiant and Cray EX AI servers with Cray HPC heritage, liquid cooling, and GreenLake services; competes with Aivres in enterprise and sovereign AI deals as a broad incumbent.
• Lenovo Group: Provides ThinkSystem SR675 V3/SR680 V3 AI servers and Neptune liquid cooling to hyperscalers and enterprises globally, with a broader PC and edge portfolio than Aivres.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• Facebookfacebook.com/AivresSystems
• LinkedInlinkedin.com/company/aivres

Revenue estimate

Valuation estimate

Number of profiles

Profiles4 records

Funding overview

Funding rounds

Investors

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 Aivres