From power
to AI.
Nachster AI sources its own renewable power and its own sites, then builds, owns and operates AI data centres in Japan. Enterprises consume that capacity three ways — a managed API, private inference endpoints, or a dedicated single-tenant cluster.
An AI infrastructure business of Nachster Energy Co., Ltd.
Compute is not the constraint.
Power, sites and operations are.
What stops an AI build in Japan is not the GPU order. It is the incoming power that carries roughly 15 kW per node, the site and grid connection that can hold it, and the team that keeps it running 24×7. Nachster AI brings all three.
Power
The sourcing and grid know-how we built investing in solar plants and battery storage stations now points at AI data centres. We contract the power before a delivery window is offered — renewable generation and grid supply together.
Data centres & power → 02 / SITESSites
We run a real-estate practice, so site acquisition and local consent are in-house work, not a consultant's deliverable. Grid, water, road access and neighbours are all secured as one package.
If you hold land → 03 / OPERATIONSOperations
A cluster only becomes capacity once it passes acceptance and runs 24×7. We deliver against the Nachster AI acceptance standard and operate with [our operating partner].
Operations & SLA →Private compute, at the granularity you need.
Capacity is allocated by window, not sold off a shelf. A Capacity Reservation fixes your allocation and your delivery date.
Managed API
An OpenAI-compatible endpoint. Billed per token, from a single API key. Models are offered by capability class.
Private endpoints
A private inference endpoint we run for you on single-tenant nodes. Billed per reserved GPU, from one node of 8 × B300.
Dedicated clusters
Single-tenant bare metal — you run your own stack. From 32 nodes (256 GPUs) to 256 nodes (2,048 GPUs).
Nachster AI B300 — standard node and first cluster
| Platform | HGX B300 NVL16 · 8-GPU node · 10U air-cooled |
|---|---|
| GPU | 8 × NVIDIA B300 · 288 GB HBM3e each · ~2.3 TB / node |
| CPU · memory | 2 × Intel Xeon 6 · 3 TB DDR5-6400 |
| Local storage | 2 × 960 GB M.2 (OS) · 8 × 3.84 TB NVMe |
| Networking | 8 × ConnectX-8 800G (integrated) · BlueField-3 DPU |
| Cluster · fabric | 32 nodes · 256 GPUs · Quantum-X800 InfiniBand · rail-optimized · 1:1 |
| Shared storage | Dedicated high-performance storage |
| Power | ~15 kW per node (reference) |
| Location | Japan · dedicated, single tenant |
| Operations · SLA | Managed 24×7 · ≥99.5% per allocation unit / month |
| Contract | Multi-year committed capacity · 5-year standard |
Standard node configuration; the final BOM is attached to the allocation offer. Later generations — GB300 · Vera Rubin — are allocated on subsequent windows.
From the substation to the GPU — one operator.
Incoming power, cooling, nodes, fabric and operations are designed as one system. We are neither handing you bare GPUs nor brokering someone else's assets.
- Monitoring and first response
- Hardware fault triage
- Spare parts held on site
- Escalation to Nachster AI
Measured per allocation unit and per cluster. Run 24×7.
A named 24×7 NOC, escalating to Nachster AI. Availability is measured at both the allocation-unit and the cluster level.
Five steps. Qualified in 48 hours. The offer stands 14 days.
Enquiry
GPU type, count, required date, workload, contract term, location and networking requirements.
Qualification
NDA, then a technical meeting with the Nachster AI platform team.
Allocation offer
A commercial term sheet with a fixed delivery window.
Contract & reservation
Agreement signed and Capacity Reservation paid; allocation and window are locked.
Build → acceptance → in service
Deployment, burn-in and acceptance testing against the Nachster AI acceptance standard.
Connecting power and land to AI demand.
We are continuously looking for AI data centre sites and power in Japan. If you hold land, generation or storage, operate a data centre, or invest in infrastructure, please get in touch.
Tell us what you need. We reply within 48 hours.
Qualified enquiries receive proposed times for a technical meeting within 48 hours, and an allocation offer after that meeting. Information is handled under NDA.