Comino

What Comino does

Comino is a London-headquartered, privately held manufacturer of liquid-cooled GPU computing infrastructure, founded in 2017 and operating across the UK, Latvia, USA, and China with 51–100 employees. The company's core technology is a proprietary closed-loop liquid cooling system that simultaneously cools GPU, VRAM, and VRM components, paired with TheQ quick-disconnect couplings that allow component swaps without draining coolant. Its flagship Grando product line comprises silent multi-GPU workstations (up to 4 GPUs) and performant multi-GPU servers (up to 8 GPUs) supporting the latest NVIDIA (H100, H200, RTX PRO 6000 Blackwell, RTX 5090, DGX Spark) and AMD (EPYC, Radeon PRO W7900) hardware, complemented by a component ecosystem including custom copper waterblocks, an Integration Kit for retrofitting air-cooled servers, Comino Energia PSUs, and a Grando mobile data center.

Comino monetizes primarily through hardware sales — Grando systems priced on request, GPU waterblocks (€800+), integration kits, and PSUs — supplemented by extended warranty subscriptions (5% per year for years 3–5), managed FPGA co-location services (€6,365–€6,480/year in a Sweden data center), and OEM/ODM design services. It also sells the Comino Cosmos AI Assistant software platform, an agentic 'AI Employees' product launched in 2026. Distribution is hybrid: direct online sales through comino.com plus an online Grando configurator, alongside a channel network spanning Puget Systems (US), PNY Technologies (EMEA), mh-Service (Germany), Teel Technologies Canada (forensics), and a strategic ModulEdge partnership for turnkey modular data center deployments across Europe and MENA. Customers span AI/ML organizations, digital forensics/law enforcement, CG and virtual production studios (e.g., powering Netflix's Motorvalley via Luno Studios), life science research institutions, and high-frequency trading firms.

Comino holds AMD Partner Program Elite status and partners directly with NVIDIA, with active presence at COMPUTEX Taipei, NVIDIA GTC, Hannover Messe, ISS World Europe, and GPEC. The company has raised approximately $1.18M via Seedrs (December 2020), its only disclosed funding round, and is targeting the projected market shift toward liquid cooling adoption in AI servers (cited as moving from 15% in 2024 to 76% by end of 2026 by Goldman Sachs).

Name

Comino

Legal name

COMINO HOLDING LTD.

Website

https://comino.com

Company type

Private

Founded year

2017

Operating status

Operating

Headcount range

51–100 employees

Short description

Comino designs and manufactures proprietary liquid-cooled multi-GPU workstations and servers (Grando brand) with up to 8 GPUs, serving AI/ML, digital forensics, virtual production, life science, and HFT customers globally through direct sales, channel partners, and OEM services.

Ownership category

akta.pro rank

Product category

High-performance liquid-cooled GPU computing hardware

NAICS

Air-Conditioning and Warm Air Heating Equipment and Commercial and Industrial Refrigeration Equipment Manufacturing (333415), Computer Systems Design and Related Services (5415)

SIC

Refrigeration & Service Industry Machinery (3580)

akta.pro primary industry

GPU-Accelerated & AI Training/Inference Servers (HDACABAG)

akta.pro secondary industries

Cooling Systems & Thermal Management (CRAC/CRAH, Chillers, Liquid Cooling) (HDABAKAD), HPC Cooling, Power & Infrastructure (Liquid Cooling, Racks, Power Delivery) (HDACAOAI), AI Datacenter Power, Cooling & Racks (liquid cooling, power delivery) (HDAAAAAH), GPU/Accelerator Compute Servers (HDABADAC), Cooling & Thermal Solutions (Air/Liquid, fans, thermal paste) (HDACAGAF), Graphics Processing Units (GPUs) & Parallel Accelerators (HDAHAJAE)

Keywords

• Liquid-cooled GPU servers
• Multi-GPU workstations
• AI infrastructure hardware
• High-performance computing systems
• Data center liquid cooling

Headquarters

HQ city

London

HQ country

United Kingdom

HQ region

Europe

Offices5 records

Markets served

GTM type

B2B

Offering type

Hardware or Manufacturing

Cost components

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

Revenue model

1. Hardware Sales: Comino generates revenue primarily from selling liquid-cooled GPU workstations, servers, and multi-GPU computing systems. Products include Grando workstations (up to 8 GPUs), integration kits for upgrading existing servers, and liquid cooling components including waterblocks, drycoolers, and power supplies. Hardware is sold directly and through channel partners.
2. Extended Warranty Services: Comino offers extended warranty programs covering 3rd, 4th, and 5th year extensions at 5% incremental cost per year. Includes remote testing, maintenance, and optional on-site deployment services with spare parts reservation.
3. FPGA Co-location Services: BittWare CVP-13 FPGA boards available with 1-year co-location in Sweden data center, including pre-installed liquid cooling, electricity, maintenance, and unlimited bitstream access.

Pricing tiers

Go-to-market motion3 records

Distribution channels6 records

Marketing channels7 records

Core offering

Comino designs and manufactures liquid-cooled multi-GPU workstations and servers under the Grando brand, supporting up to 8 NVIDIA or AMD GPUs (including H100, H200, A100, RTX PRO 6000 Blackwell, RTX 5090, and Radeon PRO W7900) in a 4U rackable form factor. The proprietary closed-loop liquid cooling simultaneously cools GPU, VRAM, and VRM, allowing sustained 24/7 operation at boost clocks with PUE as low as 1.05. Complementary offerings include GPU waterblocks, the TheQ quick-disconnect coupling system, a Liquid Cooling Integration Kit for upgrading air-cooled servers, redundant CRPS power supplies, the OTTO gaming station, and the Comino Cosmos AI assistant software.

Product overview

Comino is a European liquid cooling technology company specializing in high-performance GPU computing infrastructure. The product portfolio centers on the Grando line of liquid-cooled multi-GPU workstations and servers (Grando Workstation with up to 2x/4x GPUs, Grando Server with up to 8x GPUs), complemented by an extensive component ecosystem including GPU waterblocks for NVIDIA (H100, H200, RTX PRO 6000, RTX 6000 Ada, L40S, A100, RTX 4090, RTX 5090) and AMD (Radeon PRO W7900) GPUs, the Liquid Cooling Integration Kit for upgrading air-cooled GPU servers, Quick Disconnect Couplings (TheQ) for maintenance-free component replacement, and Comino Energia power supply units (up to 4800W with 94% efficiency). Supporting products include the BittWare CVP-13 FPGA co-location service, OTTO Gaming Station, legacy Comino N1 and A1 crypto/AI accelerators, and Comino Cosmos AI Assistant software platform. The company operates an OEM/ODM service for custom hardware solutions and offers Grando configurations optimized for specific verticals including AI inference/training, virtual production, and digital forensics. Production facilities are located in Germany with global distribution.

Differentiator

Problem solved

Functional benefit

Brands

• Grando: Hyper performance liquid-cooled multi-GPU workstations and servers available in silent workstation (up to 4x GPUs) and performant server (up to 8x GPUs) configurations.
• OTTO
• Comino Cosmos
• Comino Energia

Products and services

• Grando Workstation (Silent Workstation) 4U desktop/rackable liquid-cooled multi-GPU silent workstation supporting up to 2x NVIDIA RTX 5090 or 4x GPUs for AI inference, rendering, and scientific computing workloads.
• Grando Server (Performant Server) Liquid-cooled multi-GPU performant server with redundant power supplies, supporting 4x, 6x, or 8x NVIDIA RTX PRO 6000 Blackwell GPUs per compute node, designed for enterprise AI deployment requiring high GPU density.
• Grando Mobile Data Center Portable liquid-cooled GPU computing solution for on-site AI workloads, enabling rapid deployment of GPU-accelerated computing in remote or temporary locations with full liquid cooling infrastructure.
• NVIDIA DGX Spark (Partner Device) Compact personal AI supercomputer powered by NVIDIA Grace Blackwell architecture, offered as a partner device providing data-center-class AI capabilities in a desktop form factor.
• Liquid Cooling Integration Kit for AI Multi-GPU Server Retrofits any multi-GPU air-cooled server with liquid cooling, boosting LLM training/inference performance up to 30% and lowering facility power consumption up to 40%; supports 8x NVIDIA H200, H100, L40S, or AMD W7900 GPUs and enables operation up to +40°C without thermal throttling.
• GPU Waterblock for NVIDIA H100 80GB/94GB Full-cover liquid cooling waterblock for NVIDIA H100 PCIe GPUs (80GB and 94GB variants), with copper/stainless steel/aluminum construction that increases heat dissipation up to 10x vs air cooling.
• GPU Waterblock for NVIDIA RTX PRO 6000 Full-cover copper waterblock for NVIDIA RTX PRO 6000 Blackwell Edition GPU providing liquid cooling for GPU core, GDDR memory, and VRM components.
• GPU Waterblock for NVIDIA RTX 6000 Ada / L40 / L40S Liquid cooling solution for NVIDIA RTX 6000 Ada, L40, and L40S GPUs with copper water block and stainless steel/aluminum components for GPU, GDDR, and VRM cooling.
• GPU Waterblock for NVIDIA H200 Full-cover liquid cooling solution for NVIDIA H200 GPU with copper cold plate and stainless steel construction for GPU, GDDR, and VRM cooling.
• GPU Waterblock for NVIDIA A100 80GB Liquid cooling waterblock for NVIDIA A100 Tesla 80GB GPU with aluminum backplate and POM/stainless steel thread adapters for comprehensive GPU thermal management.
• GPU Waterblock for NVIDIA RTX A6000 / A40 Liquid cooling solution for NVIDIA RTX A6000 and A40 48GB professional GPUs providing full-cover cooling for GPU, GDDR, and VRM components.
• GPU Waterblock for NVIDIA GeForce RTX 4090 Full-cover copper waterblock for NVIDIA GeForce RTX 4090 GPUs from ASUS and GIGABYTE, cooling GPU, GDDR, and VRM with single-slot PCI design and rear-positioned fittings for server integration.
• GPU Waterblock for NVIDIA GeForce RTX 5090 Liquid cooling waterblock for GIGABYTE GeForce RTX 5090 GPU with copper cold plate and stainless steel construction for comprehensive thermal management.
• GPU Waterblock for AMD Radeon PRO W7900 Full-cover liquid cooling solution for AMD Radeon PRO W7900 48GB professional GPU with copper, stainless steel, and plastic construction for GPU and VRM cooling.
• TheQ Quick Disconnect Couplings Swedish stainless steel AISI 303 quick-disconnect couplings with G1/4" connections and 5mm nominal flow, enabling tool-free component replacement without draining the liquid cooling system.
• Comino Energia PSU CRPS (1600W / 3200W / 4800W) High-efficiency (94% 80 PLUS Platinum certified) redundant power supply units for data centers with PMBus monitoring, overvoltage/overcurrent/overtemperature protection, and remote on/off control for 24/7 server operation.
• Comino Energia Power Distribution Bus Power distribution system for liquid-cooled server deployments with 6-pin (12V/72W + 5V/30W) and 8-pin (12V/216W) connectors, sharing 5V rails with the ATX connector for flexible power delivery.
• BittWare CVP-13 FPGA Computing Service (Co-location) Co-location FPGA computing service using Xilinx Virtex UltraScale+ VU13P FPGA boards with liquid cooling; supports SHA-256 (29 GH/s), Keccak (24 GH/s), and ZenProtocol algorithms, with optional 1-year co-location in Sweden including electricity and maintenance.
• OTTO Gaming Station Compact liquid-cooled gaming PC with copper full-cover water blocks cooling GPU, VRM, CPU, south bridge, and m.2 SSD; operates at 40dB office / 32dB home noise levels, dissipates ~1kW heat, and features a semi-translucent chassis with RGB lighting.
• Comino Cosmos AI Assistant AI assistant platform providing autonomous AI agents ('AI Employees') for enterprise workforce tasks, complementing Comino's hardware infrastructure with AI automation capabilities.
• OEM / ODM Services Custom hardware design and manufacturing services leveraging Comino's liquid cooling technology and EU production capabilities, offering white-label and custom integration options for enterprise and industrial applications.
• Grando for Virtual Production & CG Specialized Grando configuration for virtual production studios, optimizing LED wall and Unreal Engine workflows; operates in harsh environments up to +40°C with low noise at full load, providing stable maximum FPS and 100% boost clock operation without thermal throttling.
• Grando Forensic (DPR Series) Liquid-cooled GPU systems optimized for digital forensics, password recovery, and encrypted evidence processing; features GPU-based password cracking and hashcat-class decryption with sustained performance under heavy 24/7 workloads, targeted at law enforcement and forensic laboratories.
• Grando Configurator Online configuration tool allowing customers to customize Grando systems with specific GPU, CPU, memory, and storage options and request tailored quotes for liquid-cooled GPU computing solutions.
• Comino X16/X16 Universal Riser High-end PCIe x16 riser with Molex gold-plated connectors and ultra-thin flexible twinax cable, supporting up to 1-meter cable length with 0% signal loss, PCIe 3.0 at 16GT/s speeds, and LED indication for operating mode and connection status.

Quantifiable outcome

• Performance boost up to 30% for AI GPU servers
• +7 more outcomes

Named customers14 records

Segments5 records

Ideal customer profiles5 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability11 records

Feature5 records

Partnerships

Nine partnerships are on record, tiered core, major, strategic and minor.

• ModulEdgecoreStrategic or Co-development Partner · 24 March 2026Partnership between Comino and ModulEdge (Czech modular data center manufacturer) to deliver turnkey on-premise AI infrastructure across Europe and MENA. Combines ModulEdge's modular data centers with Comino's liquid-cooled GPU systems featuring Nvidia hardware. Reduces deployment timelines from 12-18 months to 3-6 months, improves data sovereignty, and reduces costs and carbon footprint through energy-efficient cooling.
• NVIDIAcoreTechnology or IntegrationComino is a key NVIDIA partner offering DGX Spark, H200, and RTX PRO 6000 Blackwell GPUs. The partnership strengthened at GTC 2026, with NVIDIA leadership engaging with Comino as a key partner in bringing liquid-cooled AI infrastructure to market. Comino attends NVIDIA GTC annually.
• AMDcoreTechnology or IntegrationAMD Partner Program Elite member. Comino offers AMD EPYC processors and AMD Radeon PRO GPUs in Grando systems. Provides AMD & Comino ROCm Experience demo access for customers.
• PNY TechnologiesmajorChannel Partner/ Reseller/ DistributorOne of NVIDIA's top board partners, PNY is growing collaboration with Comino in EMEA. Met at Hannover Messe 2026 to discuss continued partnership development and increasing market demand for GPU-based solutions.
• Puget SystemsmajorChannel Partner/ Reseller/ DistributorUS channel partner that has played a key role in bringing Comino's liquid-cooled servers to American customers across AI research, defense, and entertainment sectors including Lockheed Martin and leading AI startups.
• Luno StudiosstrategicTechnology or IntegrationVirtual production studio that partnered with Comino for Netflix's Motorvalley project. Leveraged Comino liquid-cooled servers for LED wall playback, handling multiple streams of high-resolution 8K media. Worked closely with Comino to shape systems addressing practical realities of film production.
• Teel Technologies CanadastrategicChannel Partner/ Reseller/ DistributorLeading provider of digital forensic training, tools, and services nationwide. Represented Comino at ISS World Europe 2026, showcasing Grando for GPU-based password cracking and forensic analysis to law enforcement, military, and government clients.
• mh-ServiceminorChannel Partner/ Reseller/ DistributorGerman partner that showcased Comino devices at GPEC 2026 for internal security, law enforcement, digital forensics, mobile solutions, and crypto investigations professionals.
• TD SynnexminorChannel Partner/ Reseller/ DistributorMajor technology distributor connected with Comino at GTC 2026 for potential partnership opportunities in bringing Comino solutions to broader market.

Scale indicators6 records

Recent moves6 records

Expansion highlights6 records

Broad incumbents

• Supermicro: Supermicro is the established specialist in multi-GPU servers using NVIDIA H100/H200/Blackwell; it sells broadly across enterprises and is rapidly adding liquid-cooled SKUs that overlap with Comino Grando Server.
• Lenovo (ThinkStation / ThinkSystem): Lenovo ThinkStation and ThinkSystem GPU servers serve AI, VFX and research customers similar to Comino's; Lenovo's Neptune liquid-cooling line competes directly with Comino's thermal-efficiency value proposition.
• Dell Technologies (PowerEdge / Precision): Dell's PowerEdge XR/XE and Precision workstation lines target the same AI/HPC and VFX workloads as Comino; Dell's emerging direct-liquid-cooled PowerEdge SKUs are the principal large-OEM threat to Comino.

Direct peers

• Submer: Submer provides immersion and direct liquid cooling solutions for hyperscale and edge data centers — a comparable data-center thermal management peer addressing the same AI/HPC heat-density problem as Comino Grando.
• Asetek: Asetek is a liquid cooling OEM serving both PC/gaming and data-center segments; its data center business competes head-to-head with Comino's rack-scale cooling approach, while its gaming roots mirror Comino's OTTO line.
• CoolIT Systems: CoolIT Systems manufactures direct liquid cooling (DLC) cold plates, manifolds and rack-level solutions for AI/HPC servers — directly comparable to Comino's waterblocks and Integration Kit portfolio for NVIDIA GPUs.
• Puget Systems: Puget Systems is a US custom workstation and server builder and an explicit Comino channel partner; it competes and co-sells with Comino in custom liquid-cooled workstations for AI and content creation.
• Green Revolution Cooling: GRC delivers single-phase immersion cooling tanks for HPC and AI workloads; competes with Comino in the data-center cooling segment targeting GPU-dense NVIDIA deployments.

Others

• Boyd / Vertiv (thermal management for data centers): Vertiv (including legacy Boyd thermal assets) supplies CDUs, chillers and rack-level cooling infrastructure for AI data centers — an adjacent enabler that competes with Comino's InRack Drycooler and turnkey ModulEdge solutions at the facility layer.

Emerging players

• BOXX Technologies: BOXX builds high-end custom workstations and rendering boxes for VFX, simulation and AI; overlaps with Comino Grando Workstation in target customers (studios, post-production, AI developers).

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• YouTubeyoutube.com/@Cominotech/videos
• Xx.com/CominoTech
• Facebookfacebook.com/cominotech
• LinkedInlinkedin.com/company/cominotech

Compliance2 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Subsidiaries1 record

Funding overview

Funding rounds1 record

Investors1 record

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 Comino