GLOBAL GPU
Reference architectures may include NVIDIA-based configurations.
Local AI. Ready to Run.

Integrated local AI and digital-twin compute appliance for private on-site deployment.
Integrated local AI and digital-twin compute appliance for private on-site deployment.
RC-1 can be configured around different compute platforms depending on application, software compatibility, power envelope and deployment requirements.
Reference architectures may include NVIDIA-based configurations.
Qualified domestic compute platforms may be configured where required.
Reference Configuration 01 is one implementation path, not the definition of RC-1.
Local AI + Digital Twin Appliance
REFERENCE CONFIGURATIONRepresentative available configuration. Final architecture depends on workload, site and project requirements.
| FIELD | VALUE | STATUS |
|---|---|---|
| CPU | Intel Xeon Gold 6430 · 32 cores / 64 threads | REFERENCE CONFIGURATION |
| Memory | 128 GB ECC · memory generation pending supplier confirmation | TBD |
| Local high-speed storage | 4 TB NVMe enterprise SSD | REFERENCE CONFIGURATION |
| Graphics / AI acceleration | Dual NVIDIA RTX 4090 | REFERENCE CONFIGURATION |
| Archive / local data storage | 10 TB enterprise HDD | REFERENCE CONFIGURATION |
| Network | 5G + Gigabit Ethernet | REFERENCE CONFIGURATION |
| Industrial interfaces | MQTT / OPC UA | REFERENCE CONFIGURATION |
| External connectivity | USB 3.0 / HDMI 2.1 / Gigabit Ethernet / camera, sensor and display integration | REFERENCE CONFIGURATION |
Memory capacity is a supplier reference. Processor, motherboard and memory compatibility must be reconciled before this bill of materials is committed.
Digital Twin · Visualization · Simulation · Industry applications · AI-assisted operation
Local AI models · Language models · AI agents · Private inference
Digital Twin Engine · Agent / reinforcement-learning engine · Simulation · Data fusion
Compute · Graphics / acceleration · Storage · Networking · Industrial interfaces
Power and heat rejection follow the selected hardware and operating environment. Confirm ventilation and electrical interfaces during configuration.
On-premise processing and private data workflows can be configured with customer-controlled network access.
Security can be evaluated as a configurable design consideration across compute, network, data, model and agent interfaces.
Selected controls and integration scope require project-specific validation. A complete AI cybersecurity platform is not included by default.
Explore Security & GovernanceAdd cameras, sensors, displays and industrial interfaces, or move to rack-scale capacity as the workload grows.
Configurable lifecycle support can include diagnostics, preventive maintenance, validated security updates, monitoring, modernization and expansion.
Lifecycle Support & EvolutionWarranty coverage, maintenance periods, service availability, response times, third-party licensing and upgrade scope are defined in the applicable project and service agreements.
Capabilities describe system integration. Reference configurations show an implementation path. Exact values follow project engineering. Supplier references remain provider statements.
PROJECT-VALIDATED is reserved for confirmed S1 customer configurations. No project-validated product values are published here.
| FIELD | VALUE | STATUS |
|---|---|---|
| CPU | Intel Xeon Gold 6430 · 32 cores / 64 threads | REFERENCE CONFIGURATION |
| Memory | 128 GB ECC · memory generation pending supplier confirmation | TBD |
| Local high-speed storage | 4 TB NVMe enterprise SSD | REFERENCE CONFIGURATION |
| Graphics / AI acceleration | Dual NVIDIA RTX 4090 | REFERENCE CONFIGURATION |
| Archive / local data storage | 10 TB enterprise HDD | REFERENCE CONFIGURATION |
| Network | 5G + Gigabit Ethernet | REFERENCE CONFIGURATION |
| Industrial interfaces | MQTT / OPC UA | REFERENCE CONFIGURATION |
| External connectivity | USB 3.0 / HDMI 2.1 / Gigabit Ethernet / camera, sensor and display integration | REFERENCE CONFIGURATION |
Available architecture does not establish stock availability or a project delivery date. RC-2 and RC-3 remain S1 system architectures pending a dedicated supplier configuration.
Tell us what must run, where it must run and how quickly you need capacity.
Configure Your AI InfrastructureRapid Compute focuses on deployable AI infrastructure.
Projects requiring full digital project modeling, industrial simulation, energy architecture or multi-domain infrastructure integration can continue through S1 SYSTEM.
Explore S1 SYSTEMS1 SYSTEM
Model the World. Build the System.