Midokura

GPU Cloud Infrastructuremidokura.com

Midokura supplies Supermicro GPU hardware in Japan and operates the AI Factory platform — a sovereign on-premises IaaS and Kubernetes stack with GPU virtualization, serving Japanese enterprises and public institutions with private AI infrastructure.

Company typePrivate
Founded2019
HeadquartersTokyo, Japan
Headcount51–100
GTM typeB2B
OfferingSoftware

What Midokura does

Midokura is a private GPU infrastructure company that sells and operates sovereign on-premises AI compute stacks for Japanese enterprises and public institutions, with an emerging European footprint through its Barcelona entity MIDOKURA IBÉRICA, S.L. (wholly-owned Japanese subsidiary MJKK operates the Tokyo headquarters and the localized midokura.jp product catalog). The firm supplies genuine Supermicro GPU servers in Japan with full inspection-to-maintenance lifecycle services and, through a Sony partnership, subscription-based financing for hardware purchases.

The company's flagship software asset is the AI Factory platform (v25.1.3), a two-layer IaaS/PaaS stack that provisions bare-metal GPU nodes via OpenStack Magnum, builds Kubernetes clusters, manages Ceph storage, and exposes cluster add-ons (Model Provider with vLLM, JupyterHub, KubeRay) through an IaaS Console and REST Operator API. Notable proprietary work includes AMD GPU virtualization on Kubernetes via HAMi (CNCF Sandbox), introducing Compute Unit masking, memory limiting, and an LD_AUDIT-based HIP interception mechanism for ROCm. The roadmap through 2026 extended the stack into agent-ready infrastructure: a vLLM-backed Model Provider, a GraphRAG knowledge graph PoC combining Neo4j, Qdrant, and PostgreSQL 16, a Raspberry Pi-based Edge AI MCP perception node, and positioning around Sovereign AI for Agents.

Commercially, Midokura runs an enterprise field-sales motion with custom-quote pricing on GPU hardware plus professional services for cluster setup, AI development environment build-out, and Vision AI foundation model deployments. Pricing is not publicly disclosed; distribution is direct, supplemented by the midokura.jp product catalog and event-driven partner engagement. No funding rounds, revenue figures, or named customers are disclosed for the current 2019-founded entity, and the company remains privately held with 51–100 employees.

Midokura firmographics

Firmographics

Name: Midokura
Legal name: MIDOKURA IBÉRICA, S.L.
Website: https://midokura.com
Company type: Private
Founded year: 2019
Operating status: Operating
Headcount range: 51–100 employees
Short description: Midokura supplies Supermicro GPU hardware in Japan and operates the AI Factory platform — a sovereign on-premises IaaS and Kubernetes stack with GPU virtualization, serving Japanese enterprises and public institutions with private AI infrastructure.
Ownership category
akta.pro rank

Midokura industry classification

Industry

Product category: GPU Cloud Infrastructure
NAICS: Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Computer and Peripheral Equipment Manufacturing (33411), Custom Computer Programming Services (541511)
SIC: Electronic Computers (3571), Services-Prepackaged Software (7372), Services-Computer Integrated Systems Design (7373)
akta.pro primary industry: End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA)
akta.pro secondary industries: AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers) (HDAAAAAG), GPU-Accelerated & AI Training/Inference Servers (HDACABAG), GPU/Accelerator Compute Servers (HDABADAC)

Keywords

GPU cloud infrastructure
AI workload platform
Private AI cloud
GPU server hosting
HPC cluster management

Where Midokura is headquartered

Location

Headquarters

HQ city: Tokyo
HQ country: Japan
HQ region: Asia

Offices2 records

Markets served

Midokura business model

Business model

GTM type: B2B
Offering type: Software
Cost components: Technology or R&D, Personnel, Infrastructure, Supply Chain, Marketing or Sales, Operations

Revenue model

GPU Hardware Sales: Sale of Supermicro GPU servers and related hardware products in Japan. Includes comprehensive service covering inspection, delivery, and maintenance of genuine Supermicro products.
GPU Cluster & AI Development Environment Setup: Setup and configuration services for GPU clusters and AI development environments, including bare metal server provisioning, Kubernetes cluster deployment, and network infrastructure configuration.
Vision AI Foundation Model Services: Vision AI foundation model services and support for deploying and managing AI models on GPU infrastructure.
Subscription Payment Services: Subscription-based payment services offered in partnership with Sony, enabling flexible financing options for GPU infrastructure purchases.

Pricing tiers

Model	Billing	Price
Unit Pricing	Pay-as-you-go	GPU Server Products - Custom Quote Required

Go-to-market motion1 record

Distribution channels3 records

Marketing channels5 records

Midokura product offering

Product offering

Core offering

Midokura provides GPU server hardware (Supermicro systems with NVIDIA and AMD Instinct accelerators) alongside its proprietary AI Factory platform, an IaaS/PaaS environment for deploying, orchestrating, and managing GPU resources for AI and HPC workloads. The company also delivers professional setup and integration services for GPU clusters and AI development environments, plus vision AI foundation model offerings accessible through the platform.

Product overview

Midokura is a GPU infrastructure company offering a unified platform combining hardware supply (Supermicro GPU servers) with software infrastructure (AI Factory platform). The AI Factory provides a two-layer architecture: IaaS layer for bare-metal GPU provisioning via OpenStack Magnum with Ceph storage, and PaaS layer for Kubernetes-based AI workloads with cluster add-ons. Key products include: GPU Service (hardware supply), AI Factory (software platform v25.1.3), and cluster add-ons (Model Provider with vLLM, JupyterHub, KubeRay). The platform is designed for sovereign AI deployments with VPN-based secure access, SSO authentication (Azure, Google), and support for heterogeneous GPU environments (NVIDIA, AMD). Midokura operates offices in Tokyo and Barcelona, serving enterprises and public institutions requiring full data sovereignty.

Differentiator

Problem solved

Functional benefit

Brands

AI Factory: High-end GPU servers for HPC and AI workloads delivered as a private AI factory IaaS platform

Products and services

AI Factory IaaS/PaaS platform for deploying and managing GPU resources across AI/HPC workloads, with an IaaS Console UI and Operator API for multi-tenant cluster orchestration. Targeted at enterprises and public institutions building private AI clouds.
GPU Service Curated Supermicro GPU server hardware offering with both NVIDIA and AMD Instinct accelerators, supplied to enterprises for on-premises AI/HPC deployments.
GPU Cluster & AI Development Environment Setup Service Professional service for designing, deploying, and integrating GPU clusters and AI development environments, including supporting services such as extended warranty and component replacement.
Vision AI Foundation Model Vision AI foundation model offering accessible through the Midokura platform, providing computer-vision inference capability to enterprise customers.

Quantifiable outcome

GPU scaling efficiency of approximately 93% when using CU masking on AMD GPUs
+1 more outcomes

Companies that use Midokura

Customer profile

Segments2 records

Ideal customer profiles1 record

Midokura technology and API

Technology

Technology focussed Yes

API detail

Has API: Yes
API docs
API detail

Core technology

AI maturity

App detail

Integration12 records

AI capability10 records

Feature6 records

Midokura partnerships and signals

Strategic signal

Partnerships

Four partnerships are on record, tiered core, supporting and strategic.

SupermicrocoreOthersSupermicro is the primary hardware supplier for Midokura's GPU server business. Midokura provides genuine Supermicro products known for low cost and high-efficiency cooling systems, covering the full lifecycle from inspection through delivery to maintenance.
SonysupportingOthersMidokura offers subscription-based payment services through a partnership with Sony, enabling customers to finance GPU infrastructure purchases through flexible payment options.
HAMi / CNCFstrategicTechnology or IntegrationHAMi (Heterogeneous AI Computing Virtualization Middleware) is a CNCF Sandbox project that Midokura has contributed to, implementing AMD GPU virtualization support. Midokura plans to contribute AMD GPU support upstream to the HAMi project.
GitHubsupportingTechnology or IntegrationMidokura uses GitHub for open source code distribution and documentation hosting. AI Factory documentation source code available at github.com/midokura/phoenix-documentation, along with other Midokura repositories.

Scale indicators2 records

Recent moves6 records

Expansion highlights6 records

Midokura competitors and assessment

Company assessment

Broad incumbents

NVIDIA DGX (NVIDIA): NVIDIA's enterprise AI systems (DGX/HGX) plus BaseCommand/NGC software. Comparable as the dominant full-stack private AI infrastructure offering and the primary benchmark for sovereign on-prem deployments.
Supermicro: Global server manufacturer and Midokura's primary hardware supplier. Overlaps as a provider of GPU server systems to enterprises and could compete with bundled software offerings of its own.

Direct peers

RunPod: GPU cloud platform offering on-demand and dedicated GPU instances with Kubernetes-based serverless inference. Comparable GPU compute offering with developer-friendly platform layer similar to Midokura's IaaS/PaaS stack.
CoreWeave: Specialized GPU cloud provider offering NVIDIA and now diverse accelerators for AI training/inference. Most direct competitor in the private-GPU-infrastructure category Midokura targets, with similar Kubernetes-native orchestration philosophy.
Lambda Labs: GPU cloud and on-prem GPU systems provider selling hardware and hosted compute for AI workloads. Closely comparable full-stack model combining Supermicro-class hardware with managed cluster software targeting enterprise AI buyers.
Paperspace (DigitalOcean): GPU cloud platform offering GPU virtual machines and notebooks for AI/ML workloads. Comparable as a managed GPU compute platform targeting developers and enterprises needing on-demand GPU infrastructure.
Crusoe Energy: Provider of large-scale GPU clusters and AI cloud infrastructure. Comparable in selling turnkey GPU infrastructure with software orchestration for AI training/inference at enterprise scale.

Emerging players

Vast.ai: Marketplace for peer-to-peer and third-party GPU compute. Comparable AI-focused GPU provisioning layer but with a decentralized rental model versus Midokura's private on-prem delivery.
Tensorwave: GPU cloud provider purpose-built for AI workloads using AMD Instinct accelerators. Directly comparable in betting on AMD GPU virtualization for inference and training—overlaps tightly with Midokura's AMD/HAMi focus.

Regional players

Sakura Internet (SAKURA): Japanese hosting and dedicated server provider with growing GPU cloud services for AI workloads. Comparable as a regional Japanese GPU infrastructure player but more focused on hosted cloud than private on-prem delivery.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks6 records

Key highlights7 records

Customer concentration

Midokura compliance and trust

Trust signal

Compliance1 record

Midokura financial estimates

Financial estimate

Revenue estimate

Valuation estimate

Midokura leadership team

Management profile

Number of profiles

Profiles1 record

Midokura subsidiaries and ownership

Company hierarchy

Subsidiaries1 record

Midokura funding detail

Funding detail

Funding overview

Funding rounds3 records

Investors6 records

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

Midokura M&A and investment

M&A and investment

M&A

Investments

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

Frequently asked questions about Midokura

What does Midokura do?

Is Midokura a public or private company?

Midokura is a private company. It is classified as unknown and is currently operating.

When was Midokura founded?

Midokura was founded in 2019. It employs 51 to 100 people.

Where is Midokura based?

Midokura is headquartered in Tokyo, Japan, in the Asia region.

How does Midokura make money?

Four revenue lines are on record. GPU Hardware Sales are the primary driver. The others are GPU Cluster & AI Development Environment Setup, vision AI Foundation Model Services and subscription Payment Services.

Who are Midokura's main competitors?

Broad incumbents on record are NVIDIA DGX (NVIDIA) and Supermicro. Direct peers are RunPod, CoreWeave, Lambda Labs, Paperspace (DigitalOcean) and Crusoe Energy. Emerging players are Vast.ai and Tensorwave. Sakura Internet (SAKURA) is listed as a regional player.

Does Midokura have an API?

Yes. IaaS Operator API provides REST endpoints for managing users, tenants, Kubernetes clusters, and cluster add-ons. The API uses Bearer token authentication with 'user' scope. Endpoints include: POST /api/clusters/{cluster_id}/addons (install add-on), GET /api/clusters/{cluster_id}/addons (list installed add-ons), GET /api/clusters/{cluster_id}/addons/{name} (get add-on details), DELETE /api/clusters/{cluster_id}/addons/{name} (uninstall add-on), GET /api/addons/catalog (list available add-ons), user and tenant management endpoints. Developer documentation is at docs.midokura.com/docs/service-operator/OPERATOR_API_GUIDE.

What industry is Midokura in?

Midokura's product category is GPU Cloud Infrastructure. Its primary akta.pro industry code is HDAEANAA, End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management), with a secondary code of HDAAAAAG, AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers). Its NAICS code is 51821 and its SIC code is 3571.