AI compute server series
Dense GPU nodes with front-to-back service access and optional liquid-cooled cold plates.
- Form factor
- 4U / 8U rack
- Cooling
- Air or cold-plate
- Service
- Front and rear access
Products & services
Hardware, platforms and the people who run them
Specifications for procurement, architecture for technical review, and an operating model for the years after commissioning.
- Compute & infrastructure
- Software platforms
- Professional services
Overview
Products & services
03 × 12
01
Compute & infrastructure
Hardware specified for ambient conditions, service access and a ten-year operating horizon.
Cold-plate CDU and manifold
Loop-level heat exchange with leak detection and per-rack flow visibility.
- Capacity
- Scalable in 200 kW blocks
- Monitoring
- Flow, pressure, leak
- Redundancy
- N+1 pumps
Modular data hall
Power and cooling trains pre-integrated so phase one can open before full grid capacity.
- Deployment
- Block-based
- Power
- Modular UPS and busway
- Metering
- Row and loop level
Distributed storage node
Scale-out storage for meter data, billing archives and operational telemetry.
- Topology
- Scale-out
- Protection
- Replica and erasure coding
- Use case
- Meter and archive data
02
Software platforms
Platforms that have to stay correct when millions of accounts, meters and invoices pass through them.
LS-EnergyPack billing suite
Prepaid and postpaid billing, vending, payment adapters and exception queues.
- Billing
- Prepaid + postpaid
- Payments
- Bank, wallet, agent
- Audit
- Full invoice trace
RMS / CIS utility core
Customer records, contract lifecycle and field service for water, gas and electricity.
- Coverage
- Electricity, water, gas
- Workflow
- Field service integration
- Reporting
- Regulatory templates
SEMS energy management
Hall, microgrid and campus energy in one data model with tariff-aware scheduling.
- Scope
- Hall, microgrid, campus
- Scheduling
- Tariff and forecast aware
- Carbon
- Measured ledger
AI compute orchestration
Capacity allocation, job scheduling and tenant reporting across GPU pools.
- Allocation
- Tenant quotas
- Scheduling
- Queue and priority
- Billing
- Metered usage
03
Professional services
The work that decides whether a system is still trusted three years after handover.
Planning and feasibility
Load, climate and grid assessment that produces a reference architecture and a phased budget.
- Output
- Reference architecture
- Duration
- 4–8 weeks typical
System integration
Hardware, software and third-party systems integrated and tested against the operating scenario.
- Scope
- Multi-vendor
- Testing
- Scenario-based
Localised deployment
Documentation, training and interface localisation delivered in the languages the site uses.
- Languages
- EN, ID, VI, TH, ZH
- Training
- Operator and admin
24/7 operations support
Defined response model with escalation paths agreed before commissioning.
- Coverage
- Regional on-call
- Model
- Tiered escalation
Delivery path
Delivery path
Specifications for procurement, architecture for technical review, and an operating model for the years after commissioning.
01
Site and load study
Climate data, grid quality and load profile are modelled before equipment is specified.
02
Phased hall build
Cooling and power modules commissioned in blocks that match the arriving compute.
03
Energy operations
Metering, tariff logic and maintenance calendars handed to a named operating team.
How the two engines meet
Regional structure
Delivery is run from regional nodes, with engineering and local operations staffed close to the project.
Bring us your load profile
Share the site, the load and the constraint. We will come back with an architecture and a delivery path.