Amberware Trusted AI Inference Server T24-A1
Purpose-built for the Amberwire Network, the T24 provides secure local AI inference, intelligent job processing, model routing, and verified compute services. Powered by Amberwire Server OS, it connects directly to Amberwire OS, the Amberware Inference Gateway, and Amberwire Compute Mesh to process private AI workloads across the network.
SKU: AW-T24-CUDA-A1
Model: Amberware T24-A1
GPU class: 24GB dedicated AI memory
Operating system: Amberwire Server OS
Status: Pre-production
Component | T24-A1 specification |
Processor | AMD Ryzen 9 7900/7900X, 12 cores and 24 threads |
GPU | NVIDIA GeForce RTX 3090 |
GPU memory | 24GB contiguous GDDR6X |
CUDA cores | 10,496 |
Memory bandwidth | Approximately 936 GB/s |
System memory | 96GB standard; 128GB optional DDR5 |
OS storage | 1TB NVMe SSD |
AI model storage | 4TB NVMe SSD |
Networking | 2.5Gb Ethernet; optional secondary Ethernet |
Security | TPM 2.0, UEFI Secure Boot and encrypted storage |
Power supply | 1,000W, 80 Plus Gold or Platinum |
Power protection | Recommended 1,500VA pure sine-wave UPS |
Cooling | High-airflow filtered chassis with monitored PWM fans |
Enclosure | Lockable/tamper-evident black steel tower |
Expansion | Additional NVMe, high-speed network interface or future accelerator |
