Products

Private AI,
delivered to your door.

Intel NUC workstations pre-loaded with local AI models, hardened security, and everything you need to run AI without the cloud.

Software Stack

Pre-configured AI ecosystem

Every framework installed, configured, and tested. Unbox and start building.

AI Runtime

Local LLM inference engine. Run Llama 3, Mistral, CodeLlama, Phi, Gemma, and more — all completely offline with zero telemetry.

Agent Framework

Multi-agent orchestration framework. Build teams of AI agents that collaborate, delegate, and solve complex tasks autonomously.

Development

Visual interface for model management, fine-tuning, and experimentation. Compare outputs, adjust parameters, and build custom workflows.

Agent Framework

Open-source AI agent framework for building, deploying, and managing intelligent autonomous agents on your local machine.

Hardware

NVIDIA Blackwell Ultra — GB300 NVL72 & HGX B300

ArchitectureNVIDIA Blackwell Ultra

Blackwell Ultra GPUsGB300 NVL72: 72 GPUs + 36 Grace CPUs · HGX B300: 8 GPUs

CPU Cores (NVL72)2,592 Arm Neoverse V2 Cores

Total FP4 Tensor CoreNVL72: 1,440 PFLOPS (Sparse) | 1,080 PFLOPS (Dense) · HGX B300: 144 PFLOPS | 108 PFLOPS

Total FP8/FP6 Tensor CoreNVL72: 720 PFLOPS · HGX B300: 72 PFLOPS

Total Fast MemoryNVL72: 37 TB · HGX B300: 2.1 TB

Total Memory BandwidthNVL72: 576 TB/s · HGX B300: 62 TB/s

Total NVLink Switch BandwidthNVL72: 130 TB/s · HGX B300: 14.4 TB/s

Tensor Core specs shown as Sparse | Dense. Dense is ½ sparse spec. NVDEC decoder speed-ups vs H100: 2× H.264, 1.25× HEVC, 1.25× VP9, AV1 supported.