Doku AssuranceDoku SovereignAn offering from Doku Assurance
日本語EN
Request a Japan architecture review

DOKU SOVEREIGN · JAPAN

Accepting Japan design partners

Turn Japanese GPU infrastructure into an operable sovereign AI cloud.

Doku designs and delivers an operating control layer for new sovereign AI-cloud buildouts. It brings GPU-capacity orchestration, tenant isolation and governance policy, reliability automation, usage metering, and the path from commissioning to production operations together under the infrastructure operator’s authority.

Request a Japan architecture review
Moves to a concise non-confidential inquiry form. Email remains available as a fallback.
Illustrative reference architecture; deployment scope is defined with each design partner.

Founder experience across

Cornell UniversityCornell University
Google BrainGoogle Brain
Meta AIMeta AI
RunPodRunPod
FluidstackFluidstack
Prior roles and education only. No current affiliation, customer relationship, partnership, or endorsement is implied.

The operating gap

GPU capacity is not yet a cloud.

Accelerators, power, and energized racks create infrastructure. They do not automatically create a service that can isolate tenants, recover capacity, enforce policy, and explain what customers consume.

01

Fragmented orchestration

Schedulers, cluster managers, network domains, and facility systems can expose separate operating views across accelerator types and locations.

02

Uncertain billable capacity

Installed capacity is not the same as allocatable, workload-qualified capacity. Operators need a consistent view from nameplate inventory to customer-ready supply.

03

Reliability that does not scale

Manual triage, node recovery, maintenance coordination, and failure-domain decisions can become bottlenecks as fleets and tenant commitments grow.

04

Missing commercial controls

A cloud needs tenant isolation, policy enforcement, usage evidence, allocation records, and traceable operating history—not only powered racks.

Design-partner platform scope

An operating layer from rack acceptance to tenant operations.

Doku Sovereign is being designed as a modular operating control layer above operator-owned infrastructure and existing systems. Software scope, integration method, and production sequencing are agreed for each engagement.

01PLACE

GPU orchestration

Organize queues, quotas, topology, workload placement, and operating rules across heterogeneous accelerators.

02RECOVER

Fleet reliability

Structure health signals, draining, recovery, failure domains, and maintenance-capacity workflows.

03GOVERN

Sovereignty and governance

Define customer-controlled deployment, identity and policy enforcement, audit evidence, tenant isolation, and in-country data boundaries.

04METER

Commercial operations

Define evidence and interface points for usage, allocation, service catalog, billing, and reconciliation.

05ASSURE

Commissioning and acceptance

Structure infrastructure validation, burn-in, acceptance criteria, evidence collection, and production-readiness decisions.

06OPTIMIZE

Infrastructure economics

Connect utilization, available capacity, revenue-capacity, and operating risk to operator decisions.

Reference architecture

The operating layer between GPU capacity and cloud operations.

Doku is conceived as a neutral layer above the physical estate and below the service catalog, coordinating acceptance, placement, reliability, governance, and metering workflows.

This logical model is a starting point for design discussions. It does not represent a fixed product scope or production guarantee.
Download the high-resolution PDF

JAPAN-SPECIFIC OPERATING MODE

Federate capacity without surrendering operator control.

A sovereign cloud in Japan may need to coordinate capacity distributed across facilities, GPU generations, and operating organizations without consolidating ownership. Doku’s reference model is intended to coordinate capacity discovery, policy- and topology-aware placement, maintenance and failure routing, and utilization evidence through a neutral layer while each participant retains its authority.

Illustrative Japan-specific operating model. Participation, authority, data exchange, and service responsibilities are defined with each design partner.

Integration with the existing estate

Works above the systems you already operate.

Doku Sovereign is not a monolith intended to replace schedulers, cluster managers, identity, telemetry, facility controls, billing, or the service catalog. It is designed to preserve chosen systems and connect agreed operating workflows and evidence modularly. The labels below are neutral interface domains, not claims of certified integration.

SchedulersKubernetes / SlurmIdentityTelemetryDCIM / BMSBillingService catalog
OPERATOR

The operator retains

  • Infrastructure and facility ownership
  • Identity root and key control
  • Data and tenant policy
  • Customer contracts and pricing
  • Final change approval
DOKU

Doku coordinates

  • Normalized capacity inventory
  • Policy-aware placement inputs
  • Reliability and maintenance workflows
  • Metering and allocation evidence
  • Commissioning and acceptance criteria

Deployment model

Your infrastructure. Your data. Your operating authority.

Doku is not a GPU reseller or a requirement to depend on a foreign public cloud. These are reference options for design discussion; delivery scope, operating responsibility, and support access are defined before engagement.

  1. 01

    Design and integrate

    Map hardware, facilities, networking, workloads, governance requirements, and the commercial model.

  2. 02

    Commission and launch

    Agree infrastructure validation, acceptance criteria, control-layer configuration, and the initial operating workflow.

  3. 03

    Operate and expand

    Use reliability, utilization, allocation, and tenant-operating evidence to plan phased expansion.

CUSTOMER CONTROLLED

Software deployment targeted to operator-controlled infrastructure in Japan

OPERATOR SPECIFIC

A dedicated configuration isolated for the operator

MODULAR SCOPE

Software modules with scoped integration and commissioning support

OPERATOR CONTROL BOUNDARY

Define the operating boundary before implementation.

This table identifies the dimensions reviewed in a Japan reference design. These are design targets to define in technical review and contract, not pre-existing guarantees.

Deployment location

Defined per engagement, with operator-controlled Japanese infrastructure as the reference target

Design target
Data location

Workload and tenant data remains in operator-selected systems; data paths are agreed during design

Design target
Identity and key ownership

Operator ownership is the reference target; connection to existing identity and key management is confirmed per stack

Design target
Policy

The operator owns policy and final authority; the Doku layer coordinates agreed enforcement workflows

Responsibility boundary
Evidence retention

Location, retention period, and access are defined against operator requirements

Design target
Support access

An approved, time-bounded access model is the reference target; the runbook is defined per engagement

Design target
Upgrade approval

Designed to follow the operator’s change-control and approval workflow

Design target
Infrastructure ownership

Retained by the operator or its nominated infrastructure owner

Operator controlled
Pricing and customer relationships

Pricing, contracts, and customer relationships remain with the operator

Operator controlled

Design-partner work product

What a design partner receives.

Initial work is organized into implementation-oriented technical outputs tied to the operator’s existing estate and launch plan.

  1. 01Architecture map
  2. 02Sovereignty and operator boundary
  3. 03Integration plan
  4. 04Commissioning criteria
  5. 05One prioritized workflow or pilot
  6. 06Reliability, utilization, and evidence model
  7. 07Rollout and operator handoff

Useful starting points

New cloud commissioningMulti-site expansionFederated capacityTenant and commercial operationsReliability and billable evidence

Initial conversations are technical fit reviews scoped to the operator’s infrastructure, governance requirements, existing stack, and launch plan. No confidential information is required.

Request a Japan architecture review

FOUNDER-LED

Founder-led by Jorg Doku.

Jorg Doku leads Doku Sovereign design conversations directly, drawing on work across production GPU-cloud operations, frontier AI, and compiler and accelerator infrastructure.

View an illustrative report
01

Production GPU cloud

GPU-cloud deployment and operating experience at RunPod and Fluidstack.

02

Frontier AI

Research and engineering experience at Google Brain and Meta AI.

03

Systems infrastructure

Technical work spanning compilers, accelerators, and orchestration.

04

Operations and economics

A practical focus on reliability, commissioning, capacity operations, and infrastructure economics.

Prior employers are not customers, partners, or endorsers and do not imply current affiliation.

NON-CONFIDENTIAL INQUIRY

Request a Japan technical fit review.

Tell us briefly about your current environment and planning stage using non-confidential information only. We use the submission solely to assess fit and respond.