High-Density Power
Support today’s most demanding AI hardware with rack densities up to 120 kW and infrastructure designed for next-generation GPU deployments.
High-Density Colocation Built for the GPU Era.
Secure, scalable, and HPC-ready infrastructure.
Redundant Web Services (RWS) delivers dedicated, high-performance compute infrastructure specifically engineered for the most demanding AI and deep learning workloads. Our facilities are purpose-built for sustained, always-on high-performance compute (HPC).
-
Massive Power on Tap Immediate expansion capacity of 88 MW across our network.
-
Rapid Deployment Go from contract to live in 90-120 days for large-scale deployments.
-
Reliability Guaranteed 100% uptime guarantee backed by access to 8 redundant locations.
-
GPU-First Engineering Purpose built for HPC environments, capable of supporting the most power dense racks in the industry.
Built for Enterprise AI Infrastructure
Up to 120 kW per rack
Air and liquid cooling
Deploy in as little as 90 days
88 MW expansion underway
Multiple U.S. data center locations
Designed for GPU, HPC, and AI clusters
Why RWS
WHY RWS
Why Organizations Choose
RWS for AI Colocation
Faster Deployment
Avoid lengthy construction projects and utility delays. Deploy new infrastructure in as little as 90 days and accelerate time to production.
Advanced Cooling
Support demanding heat profiles with air and liquid cooling paths engineered for high-density GPU racks.
Flexible Deployment
Use tenant-owned hardware, RWS-provided hardware, or hybrid capacity with cabinet, cage, and suite configurations that scale with demand.
Enterprise Connectivity
Access redundant network providers, low-latency connectivity, and resilient infrastructure engineered for mission-critical environments.
Expert Engineering
Our infrastructure specialists work alongside your team to design power, cooling, networking, and deployment strategies tailored to your environment.
Network
RWS NETWORK
Backed by a multi-site
RWS footprint.
Place AI infrastructure near power, network paths, regional demand, and continuity requirements across the RWS site portfolio.
AI Workloads
AI WORKLOADS
Built for Today’s Most Demanding Workloads
Our facilities support a wide range of compute-intensive applications, including:
AI model training
Large language models (LLMs)
AI inference clusters
High Performance Computing (HPC)
Scientific computing
Machine learning platforms
Enterprise GPU infrastructure
Cloud and SaaS platforms
Financial modeling
Energy and engineering simulations
Location Specifications
LOCATION CAPACITY
Location capacity and
GPU specifications.
Each site is planned around power, cooling, connectivity, and GPU density assumptions that let teams compare realistic deployment footprints before committing hardware.
| Site Location | Total Power | V100 Capacity | A100 Capacity | B200 Capacity | H100, B300, GB300 |
|---|---|---|---|---|---|
| Odessa, TX | 24 MW | 16,800 | 14,400 | 12,000 | Supported |
| George West, TX | 13 MW | 9,100 | 7,800 | 6,500 | Supported |
| McAllen, TX | 10 MW | 7,000 | 6,000 | 5,000 | Supported |
| Wenatchee, WA | 6 MW | 14,000 | 3,600 | 3,000 | Supported |
| Bay City, TX | 10 MW | 7,000 | 6,000 | 5,000 | Supported |
| Current Totals | 63 MW | 53,900 | 37,800 | 31,500 | Supported |
GPU capacities are based on standard per-MW densities inside 120 kW modules: V100 (700/MW), A100 (600/MW), and B200 (500/MW).