Managed Server Colocation in Jakarta
One provider for rack space, power, connectivity, installation and ongoing server operations across multiple data-center facilities in Jakarta.
From standard 1U enterprise servers to high-density NVIDIA H100, H200, B200 and B300 infrastructure, GDI selects the appropriate facility and manages the deployment from installation to monthly reporting.
The Right Facility for the Right Server
- Standard 1U Server1U · 0.45 kW
- NVIDIA DGX H1008U · 10.2 kW
- NVIDIA DGX B20010U · 14.3 kW
- NVIDIA DGX B30010U · 14.5 kW
One provider. Multiple data centers. One point of responsibility.
Multiple Facilities
GDI can allocate infrastructure across different partner data centers.
Single Point of Contact
One GDI team coordinates facility, network and on-site operations.
GPU-Ready Planning
Power and cooling requirements are evaluated before deployment.
Operational Reporting
Regular infrastructure, incident and service reporting for the customer.
More Than Rack Space
Colocation becomes difficult when the customer has to coordinate separately with a data center, network provider, local engineer and hardware vendor. GDI provides one managed layer for the complete deployment.
Facility Selection
We match the server profile with suitable facilities based on rack capacity, power density, cooling, connectivity, security requirements and budget.
Rack & Power Allocation
GDI reserves the required rack units, power capacity, feeds and PDU configuration for the selected equipment.
Installation & Cabling
Server receiving, rack installation, power connection, network cabling, labeling and initial physical checks can be coordinated by GDI.
Connectivity
Internet bandwidth, domestic connectivity, international bandwidth, public IP, BGP and cross-connect requirements can be included in the deployment.
Remote Hands
Physical reboot, cable checks, visual inspection, component replacement and other approved on-site tasks can be coordinated without sending your own engineer to Jakarta.
Monitoring & Incidents
GDI becomes the operational coordination point for infrastructure incidents and escalates issues to the appropriate facility or network team.
Hardware Logistics
Optional receiving, inventory control, spare-parts handling, RMA coordination and hardware replacement support.
Reporting
Customers can receive recurring reports covering assets, incidents, maintenance, service activity and available infrastructure metrics.
The Right Facility for the Right Server
Not every data center is suitable for every workload.
A standard 1U application server may only require a few hundred watts, while an NVIDIA DGX B200 can require up to approximately 14.3 kW and a DGX B300 approximately 14.5 kW of operating power, with significantly higher rack-density planning requirements.
Instead of forcing every customer into one facility, GDI can select from multiple partner locations and allocate the infrastructure according to the actual workload.
Standard Enterprise
Priority: cost efficiency, reliable connectivity, normal rack density.
Business Critical
Priority: redundant power/network options, stronger SLA requirements, controlled access and reporting.
AI / High Density
Priority: high power per rack, cooling capability, electrical design, rack depth/weight, high-speed network and expansion capacity.
Server Categories
Pick the profile closest to your hardware. You do not need exact rack engineering details — presets carry sensible defaults, and Custom lets you enter your own.
Standard 1U Server
1U · 0.45 kW power profile
Single-socket or dual-socket 1U compute servers for applications, web and general workloads.
- Dell PowerEdge R660-class
- HPE ProLiant DL360-class
- Supermicro 1U compute
From
Rp 1.560.000/ month
≈ $88.6
One-time setup from Rp 660.000 (≈ $37.5)
Standard 2U Server
2U · 0.5 kW power profile
2U compute servers with more storage bays or expansion cards, within the standard power allowance.
- Dell PowerEdge R760-class
- HPE ProLiant DL380-class
- Lenovo ThinkSystem SR650-class
From
Rp 2.040.000/ month
≈ $116
One-time setup from Rp 660.000 (≈ $37.5)
Storage / High-Power 2U
2U · 1.2 kW pricing power
Storage appliances or CPU-dense systems above the standard power allowance. Priced with additional reserved power.
- Storage appliances
- CPU-dense 2U systems above the standard power allowance
From
Rp 4.980.000/ month
≈ $283
One-time setup from Rp 660.000 (≈ $37.5)
4U GPU — L40S / PCIe GPU Class
4U · 3 kW max system power
PCIe GPU servers for inference and mid-size training. OEM designs vary, so this is an estimate preset, not a guaranteed electrical specification.
- 4× NVIDIA L40S
- 4× PCIe accelerator server
- OEM AI inference server
From
Rp 13.080.000/ server / month
≈ $743
One-time setup from Rp 6.000.000 (≈ $341)
Final price depends on the actual server PSU and measured / reserved power profile.
NVIDIA DGX H100
8U · 10.2 kW max system power
8U air-cooled system with 8× H100 GPUs. High-density placement, engineering validation required.
From
Rp 36.360.000/ server / month
≈ $2,066
One-time setup from Rp 6.000.000 (≈ $341)
Facility validation required
NVIDIA DGX H200
8U · 10.2 kW max system power
Same chassis and power envelope as DGX H100 with 8× H200 GPUs.
From
Rp 36.360.000/ server / month
≈ $2,066
One-time setup from Rp 6.000.000 (≈ $341)
Facility validation required
NVIDIA DGX B200
10U · 14.3 kW max system power
NVIDIA B200 infrastructure requires data-center engineering beyond standard colocation. GDI evaluates rack layout, power feeds, cooling capacity, server weight, network design and expansion capacity before confirming the facility.
From
Rp 49.500.000/ server / month
≈ $2,813
One-time setup from Rp 6.000.000 (≈ $341)
Facility validation required
NVIDIA DGX B300
10U · ~14.5 kW typical power · 15 kW planning value
For B300 deployments, GDI validates the electrical design, rack configuration, cooling profile and network architecture with the selected facility before installation.
From
Rp 51.600.000/ server / month
≈ $2,932
One-time setup from Rp 6.000.000 (≈ $341)
Facility validation required
Custom Server
Enter rack units, power, feeds, depth, weight and ports for any other system. Above 5 kW per server it is classified as high density; above 10 kW an engineering review is required before final quotation.
Colocation for NVIDIA B200 & B300
Blackwell-class systems cannot be treated like standard rack servers. A DGX B200 occupies 10U and can require up to 14.3 kW. DGX B300 is also a 10U system and operates around the 14.5 kW class, with a 15 kW system maximum used for capacity planning.
At this density, available rack space is only one part of the problem. The facility must be able to support the required electrical design, cooling, airflow, rack weight, PDU or busbar architecture and high-speed network connectivity.
GDI evaluates these requirements across suitable partner facilities before confirming a deployment.
Rack planning
Default: 2 systems per rack
Specialized: 4 systems per rack — specialized high-density facility required, subject to engineering approval.
NVIDIA DGX B200
- 10U
- 14.3 kW max
- 8× B200 GPUs
- High-density validation required
From
Rp 49.500.000/ server / month
≈ $2,813
One-time setup from Rp 6.000.000 (≈ $341)
NVIDIA DGX B300
- 10U
- ~14.5 kW operating class
- 15 kW planning value
- 8× B300 GPUs
From
Rp 51.600.000/ server / month
≈ $2,932
One-time setup from Rp 6.000.000 (≈ $341)
Estimate Your Colocation Cost
Select your server type and quantity. We will estimate rack space, reserved power and monthly colocation cost. High-density GPU estimates are indicative and require final facility validation.
Configure
Leave empty to use the included power profile.
1–200 servers
Long-term and volume pricing may be available after facility confirmation.
OS-level management and hardware responsibility for your servers. GPU nodes use the GPU / AI Node Operations plan. Price depends on the hardware value of the selected server.
Your estimate
Estimated Monthly Colocation
Rp 1.560.000/ month
≈ $88.6 / month
1/42U · 1 servers / rack
This is a planning estimate, not a binding quotation.
Indicative prices only and subject to technical validation, facility availability and contract conditions. High-density GPU deployments require confirmation of power, cooling, rack and network capacity. Prices exclude applicable taxes and optional third-party connectivity/cross-connect charges unless specifically included in the final proposal. USD amounts are indicative conversions at Rp 17.600 per USD for reference only; invoicing is in Indonesian rupiah (IDR).
Standard Colocation — Indicative Pricing
1U Server
From
Rp 1.560.000/ month
≈ $88.6
- Up to ~450 W profile
- Basic GDI coordination
- 1 server
2U Server
From
Rp 2.040.000/ month
≈ $116
- Up to ~500 W profile
- Basic GDI coordination
- 1 server
Quarter Rack
From
Rp 6.600.000/ month
≈ $375
- Approx. 10U class
- Power allowance depends on facility
Half Rack
From
Rp 10.800.000/ month
≈ $614
- Approx. 20–22U class
- Power allowance depends on facility
Full Rack
From
Rp 24.000.000/ month
≈ $1,364
- Approx. 42–45U class
- Standard-density power only
High-density GPU power is priced separately and should not be estimated from standard full-rack pricing.
Add-ons
- Extra 100 W (standard servers): Rp 420.000 / month (≈ $23.9)
- Additional IPv4: from Rp 36.000 / month (≈ $2.0)
- Additional LAN port: Rp 180.000 / month (≈ $10.2)
- Additional standard power socket: Rp 180.000 / month (≈ $10.2)
- OS installation assistance: Rp 420.000 / event (≈ $23.9)
- Cross-connect, BGP, private circuit, IX, dark fiber, cloud connect, AI fabric: request quote
Managed Service Levels
Every level defines the responsibility boundary between GDI, the facility and your team.
GDI Core Management
Included in base estimate
- Facility selection and coordination
- Deployment coordination
- Asset registration
- Incident escalation
- Maintenance coordination
- Basic monthly service report
- Single GDI commercial contact
Managed Remote Operations
From Rp 2.400.000 (≈ $136) / month / deployment
- Everything in Core
- Priority operational coordination
- Remote-hands allowance according to contract
- Physical checks and approved interventions
- Spare-parts / RMA coordination
- Expanded monthly report
Enterprise NOC
Custom quote
- GPU clusters and multi-rack environments
- Regulated and business-critical workloads
- 24/7 escalation requirements
- Custom incident response
- Multi-site infrastructure
Managed Server Operations
Colocation keeps your hardware powered and connected. Managed Server Operations keeps the operating system, security, backups and applications healthy — and, on the Advanced and GPU plans, GDI takes responsibility for the hardware itself: if it breaks, we diagnose, repair or replace it.
Separate from data-center remote hands (physical tasks at the facility). Because GDI carries the risk of hardware failure, plan prices scale with the value of the server: each plan includes labour and tooling plus a hardware-responsibility reserve calculated as a share of the hardware value per year.
Essential Monitoring
Know before your users do.
From · Standard 1U Server
Rp 520.000
/ server / month · ≈ $29.5
Onboarding Rp 650.000 (≈ $36.9) · 60 min response
no hardware coverage
What is included
- 24/7 uptime, CPU, RAM, disk and network monitoring
- Alerting to your team (email / WhatsApp / Telegram)
- Incident escalation to facility remote hands
- Monthly availability and health report
- No administrative work, no hardware coverage
Onboarding includesRp 650.000 · 2 engineering hours
- Connection handover after rack installation
- Monitoring and alerting agents installed
- Escalation contacts and report template configured
Standard Managed OS
Patched, hardened, backed up.
From · Standard 1U Server
Rp 2.520.000
/ server / month · ≈ $143
Onboarding Rp 2.600.000 (≈ $148) · 4 admin hours / month · 240 min response
hardware-responsibility reserve: 5% of hardware value / year
What is included
- Everything in Essential
- Monthly OS patching in an agreed window (Linux / Windows)
- Security hardening: firewall, SSH / RDP, fail2ban, updates
- User and access management
- Backup job configuration and monitoring (storage separate)
- Hardware diagnostics, vendor / RMA coordination and replacement labour — parts billed at cost
- Up to 4 admin hours / month, 4-business-hour response
Onboarding includesRp 2.600.000 · 6 engineering hours
- Connection handover, IP / VLAN and remote access (SSH / RDP / IPMI)
- OS installation or clean-up (Linux / Windows), updates, time, DNS
- Security baseline: firewall, users, keys, fail2ban, antivirus where required
- Monitoring and backup agents
- Installation and initial configuration of the software you need: web server, database, runtime, Docker, mail, control panel
- Documentation of the delivered configuration
Advanced 24/7 Managed
If it breaks, it is our problem.
From · Standard 1U Server
Rp 7.640.000
/ server / month · ≈ $434
Onboarding Rp 3.900.000 (≈ $222) · 10 admin hours / month · 30 min response
hardware-responsibility reserve: 10% of hardware value / year
What is included
- Everything in Standard
- 24/7 incident response, 30-minute urgent SLA
- Full hardware responsibility: parts, labour and replacement, next business day
- Spare-parts reserve for disks, PSUs, memory and fans
- Emergency security patching outside the window
- Application stack support: web, database, containers
- Quarterly backup restore test and DR runbook
- Up to 10 admin hours / month
Onboarding includesRp 3.900.000 · 10 engineering hours
- Everything in Standard onboarding
- Application stack tuned for production: reverse proxy, database, cache, containers / Kubernetes
- High-availability or cluster setup where applicable
- Backup policy, first full backup and restore test
- DR runbook and incident escalation matrix
GPU / AI Node Operations
For H100, H200, B200 and B300 nodes.
From · 4U GPU — L40S / PCIe GPU Class
Rp 19.280.000
/ GPU node / month · ≈ $1,095
Onboarding Rp 11.700.000 (≈ $665) · 12 admin hours / month · 30 min response
hardware-responsibility reserve: 10% of hardware value / year
What is included
- 24/7 node health: GPU, HBM, NVLink, thermals via DCGM
- Full hardware responsibility: diagnostics, RMA with NVIDIA / OEM, module and node replacement coordination
- Driver, CUDA, firmware and BMC lifecycle management
- Scheduler support: Slurm, Kubernetes, container runtime
- Burn-in, benchmark baseline and regression checks
- Cluster network health checks (Ethernet / InfiniBand)
- Up to 12 engineer hours / node / month
- Volume pricing reviewed from 10 nodes
Price scales with the node value (H100 < B200 < H200 < B300). Manufacturer support contracts (e.g. NVIDIA DGX Care) can be passed through in addition.
Onboarding includesRp 11.700.000 · 24 engineering hours
- Node connection, BMC / management network, firmware baseline
- OS, NVIDIA drivers, CUDA, container toolkit, DCGM
- Scheduler / orchestration: Slurm or Kubernetes with GPU operator
- Cluster network validation (Ethernet / InfiniBand)
- Burn-in, benchmark baseline (NCCL, HPL / MLPerf-style), thermal check
- Installation of the customer's ML stack: frameworks, model serving, storage mounts
- Handover documentation and health dashboard
Monthly price by server class
Hardware value is the typical market price of the server; the exact figure is confirmed at quotation. Custom servers use the value you enter.
| Server class | Essential Monitoring | Standard Managed OS | Advanced 24/7 Managed | GPU / AI Node Operations |
|---|---|---|---|---|
Standard 1U Server ≈ $6,000 hw | Rp 520.000 ≈ $29.5 | Rp 2.520.000 ≈ $143 | Rp 7.640.000 ≈ $434 | — |
Standard 2U Server ≈ $9,000 hw | Rp 520.000 ≈ $29.5 | Rp 2.810.000 ≈ $160 | Rp 8.220.000 ≈ $467 | — |
Storage / High-Power 2U ≈ $16,000 hw | Rp 520.000 ≈ $29.5 | Rp 3.480.000 ≈ $198 | Rp 9.550.000 ≈ $543 | — |
4U GPU — L40S / PCIe GPU Class ≈ $50,000 hw | — | — | — | Rp 19.280.000 ≈ $1,095 |
NVIDIA DGX H100 ≈ $230,000 hw | — | — | — | Rp 53.600.000 ≈ $3,045 |
NVIDIA DGX H200 ≈ $315,000 hw | — | — | — | Rp 69.810.000 ≈ $3,966 |
NVIDIA DGX B200 ≈ $290,000 hw | — | — | — | Rp 65.040.000 ≈ $3,695 |
NVIDIA DGX B300 ≈ $330,000 hw | — | — | — | Rp 72.670.000 ≈ $4,129 |
Operations add-ons
- Additional admin hour: Rp 650.000 (≈ $36.9)
- Managed backup storage: Rp 325.000 per 100 GB / month · Rp 1.625.000 per TB / month
- Monthly vulnerability scan + report: Rp 975.000 / server / month
- DR restore test / drill: Rp 1.625.000 per event
- Hardware parts for Essential / Standard plans: billed at cost + coordination hours
- Windows Server / SQL Server licensing (SPLA) and manufacturer support contracts: quoted separately
- Additional software installation or migration work beyond the included hours is billed at the additional admin hour rate.
Deployment Process
- 01
Requirements
We collect server models, quantities, power, connectivity, deployment date and operational requirements.
- 02
Facility Matching
GDI evaluates suitable partner facilities and available capacity in Jakarta / Greater Jakarta.
- 03
Technical Validation
Rack, electrical, cooling, network, weight and access requirements are checked before commitment.
- 04
Commercial Proposal
GDI issues one consolidated proposal for the selected infrastructure and managed services.
- 05
Deployment
Hardware receiving, rack installation, power, cabling and connectivity are coordinated according to the approved scope.
- 06
Operations
GDI manages ongoing coordination, incidents, remote hands, reporting and future expansion.
Know What Is Happening to Your Infrastructure
GDI can consolidate operational information into a recurring infrastructure report, giving management one view of deployed assets, service activity and incidents across the selected data-center environment.
Metrics are included where data is available from the selected facility.
Example report contents
- 01Customer / project and facility reference
- 02Asset inventory, rack location / U position where contractually allowed
- 03Allocated power and available power metrics
- 04Bandwidth / traffic summary
- 05Incidents opened / closed, downtime or SLA events reported by the facility
- 06Scheduled maintenance and remote-hands activity
- 07Hardware changes and spare-parts usage
- 08Access events
- 09Open risks, recommended actions, capacity / expansion notes
One GDI Contract. Multiple Facilities.
For customers that do not want all infrastructure in one location, GDI can design deployments across separate partner data centers.
Actual geographic separation, latency, connectivity and failover architecture depend on the selected facilities and customer requirements.
Enterprise Deployments
From ten GPU nodes upward, GDI treats the project as dedicated infrastructure: reserved racks and power blocks, phased delivery, staging and spares. The calculator flags the tier automatically.
10+ GPU nodes
Enterprise Infrastructure
Dedicated racks, reserved power blocks and negotiated facility pricing after capacity request.
20+ GPU nodes
Large AI Cluster
Dedicated infrastructure engineering required.
50+ GPU nodes
AI Factory Planning
Multi-rack power, cooling, network fabric, staging, spares and phased deployment.
100+ GPU nodes
Capacity Reservation
Facility capacity and deployment schedule confirmed before commercial commitment.
Frequently Asked Questions
GDI provides managed colocation through selected partner data-center facilities. This allows us to match the facility to the technical and commercial requirements of each project rather than limiting customers to one location.
Tell Us What You Need to Deploy
Send us the server model and quantity. GDI will evaluate suitable data-center capacity in Jakarta and prepare a consolidated colocation and managed-service proposal.
Standard servers • full racks • H100/H200 • B200/B300 • multi-site deployments
Indicative prices only and subject to technical validation, facility availability and contract conditions. High-density GPU deployments require confirmation of power, cooling, rack and network capacity. Prices exclude applicable taxes and optional third-party connectivity/cross-connect charges unless specifically included in the final proposal. USD amounts are indicative conversions at Rp 17.600 per USD for reference only; invoicing is in Indonesian rupiah (IDR).