From Land to Hand: Building AI Infrastructure That Actually Works
We’ve got some exciting news to share. RAVEL and Strata Expanse, in collabortaion with Supermicro, have officially launched our first AI Center of Excellence (COE) in Ironton, Ohio—and it’s the first step in a much bigger national rollout.
This isn’t about standing up another traditional data center. It’s about rethinking how AI infrastructure gets built, powered, and delivered—faster, closer to demand, and ready for real-world AI workloads from day one.
Why AI Centers of Excellence Matter
The Ironton COE is a working model for what modern AI infrastructure should look like.
Each Center of Excellence is designed to be:
- Production-ready from day one
- Quick to deploy (think weeks, not years)
- Built specifically for AI, not retrofitted later
These centers give enterprises, public-sector teams, and research organizations a place to test, optimize, and scale AI workloads without fighting infrastructure limitations. Whether you’re training LLMs, running GenAI pipelines, deploying RAG systems, or tackling HPC and simulation workloads, the COEs are built to support it.
Gray-Space-as-a-Service: A New Way to Build AI Infrastructure
Strata Expanse is leading the charge with a new model called Gray-Space-as-a-Service, and it changes the way AI infrastructure comes to life.
Instead of waiting years for traditional data centers to catch up, this model:
- Starts with available land
- Builds dedicated power assets
- Turns them into scalable, high-performance AI environments
- Delivers compute directly to tenants—where it’s needed
Power is the biggest constraint in AI today. Gray-Space-as-a-Service addresses that by combining natural gas, deep geothermal energy and cooling, hydrogen technology, energy storage, and solar generation to deliver clean, resilient, high-density power at scale.
The result is distributed AI infrastructure that connects locations across the country—bridging the gap between land, power, and compute.
Supermicro: The Compute Foundation
The Ironton Center of Excellence is powered by Supermicro’s high-density AI systems, providing the compute backbone for serious AI workloads.
That includes:
- Liquid-cooled and air-cooled clusters
- GPU and CPU platforms optimized for LLMs, GenAI, training, and HPC
It’s flexible, modular infrastructure designed to grow as workloads evolve.
RAVEL: The Intelligence Layer
Hardware and power are only part of the story. RAVEL provides the intelligence that ties it all together.
Our Orchestrate AI platform serves as the unified control plane across the COEs, enabling:
- Multi-vendor GPU, CPU, and TPU orchestration
- Power-aware scheduling and automation
- Multi-tenant GPUaaS and AIaaS offerings
- Consistent, production-grade operations across every site
Simply put, we make sure these centers don’t just exist—they perform.
Ohio Is Just the Beginning
The Ironton COE is the first of 30 planned sites nationwide, all designed to deliver sustainable, resilient, high-performance AI infrastructure where it’s needed most.
Together with Strata Expanse and Supermicro, we’re helping enterprises, innovators, and institutions move faster—without waiting years for power or capacity to come online.
This is what Land-to-Hand AI infrastructure looks like in practice. Ironton is live—and we’re just getting started.
Useful links related to this article