AI Compute & Colocation
GPU-class AI training and inference capacity offered as colocation with private and shared models.
- High-density racks
- Private cages and suites
- Remote-hands services
RIFSOL · AI DATACENTER DIVISION
AI Datacenter Division
Compute, powered by Morocco.
High-density AI compute, colocation and HPC services — run on 50% renewable energy and connected to three continents.
AI computeColocationLiquid cooling50% renewable
Overview
Where RifSol chips meet RifSol infrastructure
The AI Datacenter division gives the vertical-integration platform its endpoint: its own compute infrastructure. High-density racks host GPU-class and CPU-class systems for AI training and inference, operated as colocation and HPC services for customers who need secure, nearshore capacity without building it themselves.
Density is the defining engineering challenge. Modern accelerators concentrate kilowatts in a single chassis, so the facility pairs direct liquid cooling with hot-aisle containment to keep silicon in its efficiency window. Power comes from Morocco’s grid mix — already 50% renewable — with the national solar and wind build-out pointing toward a carbon-neutral trajectory for the site by 2027.
Location turns the datacenter into a connectivity product. Positioned at the crossroads of Europe, Africa and the Americas — with submarine cables and the Tangier Med / Casablanca corridor nearby — the facility offers low-latency reach into three continents from one jurisdiction. Security and compliance follow the same certified-process culture as the factories, with physical security, redundant networking and auditable operations.
50%
Renewable energy
2027
Carbon-neutral site target
3
Continents served
24/7
Facility operations
Capabilities
What AI Datacenter delivers
HPC Services
Managed high-performance computing for simulation, CAE and research workloads.
- Batch scheduling
- Parallel filesystems
- Engineering support
Liquid Cooling
Direct liquid cooling keeps high-TDP accelerators efficient at rack densities air cannot serve.
- Direct-to-chip loops
- Heat-reuse ready design
- Contained hot aisles
High-Density Racks
Power-dense rack architectures with intelligent PDUs and per-unit monitoring.
- Up to kW-per-U provisioning
- Redundant power paths
- Live telemetry
Renewable Power
Operations run on 50% renewable energy today, with a carbon-neutral site target for 2027.
- Grid + renewable mix
- Solar and wind sourcing
- Efficiency-first design
Security & Compliance
Physical and network security with auditable, certified process discipline.
- Multi-layer physical security
- Redundant connectivity
- Documented operations
How it works
The flow, step by step
- 01
Capacity Planning
Rack power, cooling and network budgets are engineered per tenant before deployment.
- 02
Deployment
Racks, cabling and cooling loops are installed and commissioned against design docs.
- 03
Onboarding
Tenant systems are racked, connected and burn-in validated before handover.
- 04
24/7 Operations
Facility monitoring, remote hands and predictive maintenance keep services up.
- 05
Optimization
Cooling setpoints, load balancing and renewable sourcing are tuned continuously.
- 06
Review & Scale
Capacity and service reviews drive the next expansion phase.
Why it matters
The same electrons, the same discipline.
The datacenter consumes the chips the platform designs and builds, and it runs with the same certified-process culture as the fabs. Efficiency — per workload, per rack, per megawatt — is tracked the way yield is tracked in the fab: as a live engineering metric the whole team can see.
Where it goes
Applications & industries
AI training & inference
Dense accelerated compute for model development and production inference nearshore.
Engineering simulation
HPC capacity for CAE, semiconductor TCAD and energy-system modeling.
Nearshore hosting
EU-adjacent colocation with data-residency and latency advantages.
Work with AI Datacenter.
From engineering programs to production partnership — the AI Datacenter team is one message away.