01/
Site and utilities
Site qualification, utility coordination, power distribution, and backup generation.
TURNKEY MODULAR AI INFRASTRUCTURE
From Power-Ready Site to First Token.
RWS coordinates the facility environment required for your GPU deployment—from site, power, and fiber through modular infrastructure, cooling, integration, testing, and commissioning. You bring the compute systems and workloads.
What Turnkey Modular Includes
The Facility Environment, Delivered as One System.
Turnkey Modular brings the major facility workstreams into one coordinated delivery scope. RWS aligns the site, utility power, fiber, modular data hall, cooling, rack infrastructure, integration, and commissioning around the requirements of your deployment.
With fewer handoffs between independent vendors, schedule dependencies and acceptance requirements can be defined before work begins.
02/
Modular facility
Modular data-hall systems and the supporting physical infrastructure.
03/
Cooling
Central plant, distribution, and rack-level cooling matched to the deployment.
04/
Power train
UPS systems, PDUs, rack-level power, and associated distribution.
05/
Connectivity and integration
Fiber, meet-me-room access, cross-connects, racks, CDUs, and agreed facility-network integration.
06/
Testing and commissioning
System verification, documented acceptance testing, commissioning, and production handoff.
Responsibility Boundary
RWS Delivers the Environment. You Bring the Compute.
Turnkey Modular gives the project one coordinated facility-delivery scope without taking control of the customer's compute architecture or workloads. Final responsibilities, interfaces, and acceptance criteria are documented during engineering.
RWS coordinates
- Site qualification and delivery planning
- Utility power and site distribution
- Backup generation and fuel systems
- Fiber and meet-me-room infrastructure
- Modular data-hall infrastructure
- Central and rack-level cooling infrastructure
- UPS systems, PDUs, racks, and CDUs
- Agreed cross-connect and facility-network integration
- Installation coordination and system testing
- Commissioning, acceptance documentation, and production handoff
Customer provides
- GPUs, servers, storage, and related compute hardware
- Compute architecture and hardware specifications
- Network architecture, security policy, and workload configuration
- Software, orchestration, models, data, and applications
- Hardware procurement decisions and vendor approvals
- Workload validation and production acceptance
Qualification note: This is the standard delivery boundary. Final system ownership, operating responsibilities, performance requirements, and acceptance criteria are established for each project during engineering and contracting.
Parallel Delivery Model
Factory and Site Work Move Forward Together.
Turnkey Modular is designed to advance workstreams in parallel. While site power, civil work, and fiber are prepared, modular facility systems can be engineered and assembled off-site. The workstreams converge for installation, integration, and commissioning.
01/
Factory work
Modular systems are engineered, assembled, and prepared around the approved deployment requirements.
02/
Site work
Power, civil, fiber, backup systems, and site infrastructure advance against the same deployment plan.
03/
Integration and commissioning
The facility systems are installed, integrated, tested, and accepted as one operating environment.
Qualification note: Schedule depends on utility availability, permitting, site conditions, equipment lead times, customer hardware readiness, and final project scope.
Path to First Token
A Defined Path to First Token.
RWS connects facility planning, delivery, and acceptance through a defined project path. Each stage establishes the inputs required for the next and makes key dependencies visible earlier.
01/
Qualify
Confirm site, power, fiber, cooling, deployment schedule, hardware plan, and operating requirements.
02/
Engineer
Define the facility design, responsibility boundary, system interfaces, performance requirements, and acceptance criteria.
03/
Build in parallel
Advance site infrastructure and modular-system work against the coordinated project plan.
04/
Integrate and commission
Install and test the facility systems, document results, and resolve acceptance items.
05/
Handoff and validate
Complete the agreed production handoff so the customer can deploy its compute and validate the workload.
What “first token” means: The facility environment and agreed integration scope have been commissioned and are ready to support customer hardware deployment and workload validation. Customer compute, software, and workload readiness remain customer responsibilities unless expressly included in the project scope.
Design Around the Deployment
Design the Environment Around the Actual Deployment.
Turnkey Modular is not a generic shell with infrastructure added later. The environment is qualified and engineered around the deployment's power, density, cooling, network, physical, operating, and acceptance requirements.
Power
IT load, redundancy, distribution, rack power, and expansion requirements.
Cooling
Heat-rejection strategy, supply temperatures, flow, rack density, and CDU requirements.
Network
Carrier access, cross-connects, pathways, demarcation, and agreed facility integration.
Physical environment
Rack layout, loading, clearances, access, security, and serviceability.
Operations
Monitoring, maintenance, escalation, access, and handoff responsibilities.
Commissioning
Test procedures, documentation, acceptance thresholds, and exception resolution.
Site Qualification
Power Is Only the Starting Point.
Available power does not automatically make a site deployment-ready. RWS evaluates the broader conditions that affect feasibility, design, schedule, operation, and future expansion.
Qualification areas
- Utility capacity, delivery conditions, and schedule
- Fiber availability, diversity, and carrier access
- Civil, permitting, and physical site conditions
- Cooling-system requirements and heat-rejection constraints
- Backup generation and fuel strategy
- Security, access, serviceability, and operating requirements
- Expansion capacity and phased-growth options
Commissioning and Acceptance
Verify the Environment Before You Commit Capacity.
Acceptance should not be defined after installation. RWS establishes the required tests, documentation, performance thresholds, and responsibility for resolving exceptions during engineering—then validates the delivered environment against that agreed basis.
Review areas
- Power-system operation and redundancy behavior
- Cooling capacity, distribution, and control performance
- Network pathways and agreed facility integration
- Monitoring, alarming, and operational visibility
- Failure-response and recovery procedures
- Documentation, open-item resolution, and production handoff
Review the proposed site, infrastructure scope, system interfaces, and acceptance approach with the RWS engineering team.
Frequently Asked Questions
Turnkey Modular FAQs
What does “turnkey” mean at RWS?
RWS coordinates the infrastructure environment surrounding the customer’s compute systems, including the site, power, fiber, modular data hall, cooling, racks, integration, testing, and commissioning. Final scope is defined for each project.
Does RWS provide the GPU hardware?
The standard turnkey model assumes the customer supplies its GPU systems and workloads. RWS may help coordinate procurement partners or provide bridge capacity when appropriate, subject to availability and separate commercial terms.
Which sites support turnkey modular deployments?
Site eligibility depends on current capacity, power, fiber, cooling, development status, deployment size, and schedule. RWS confirms supported sites during qualification.
How quickly can a turnkey environment be deployed?
Timing depends on project scope, site readiness, engineering, utility conditions, equipment availability, customer hardware, and contract terms. RWS provides a project-specific schedule after qualification.
Can RWS support liquid-cooled hardware?
RWS evaluates liquid-cooling requirements as part of site and design qualification. Supported configurations and available capacity vary by site.
Can we begin with one delivery model and expand the RWS scope later?
Potential changes depend on the site, project stage, installed systems, commercial terms, and engineering requirements. The preferred future-state model should be discussed during initial planning.
Can RWS provide temporary compute while our environment is being completed?
RWS-hosted GPU or bare-metal capacity may be available for development, migration, or bridge workloads. Availability, hardware, performance, and pricing are confirmed separately.
Plan the Deployment
Bring Us the Hardware Plan and Production Target.
Share your location, power, density, cooling, hardware, and schedule requirements. RWS will help qualify the appropriate site and define a coordinated path from facility delivery through production handoff.