M3MonkeyT3chBack to Solutions

Recommended Starting Point

CPUAMD Ryzen 9000 Series, up to 16 cores

GPUUp to 2x NVIDIA RTX 5090 or RTX PRO 6000

VRAMUp to 192 GB

MemoryUp to 256 GB DDR5

Configuration Choices

Processor

AMD Ryzen 9000 options for compact local AI and workstation use.

8-Core 3.80 GHz AMD Ryzen 7 9700XEfficient entry CPU for single-GPU inference and prototyping.+$012-Core 4.40 GHz AMD Ryzen 9 9900XMore threads for preprocessing, multitasking, and CPU-side tools.+$36316-Core 4.30 GHz AMD Ryzen 9 9950XBest fit when data prep and GPU feeding matter.+$61216-Core 4.30 GHz AMD Ryzen 9 9950X3DHigher-end CPU path for mixed AI, simulation, and interactive work.+$938

Cooling

Choose between quieter liquid cooling and simpler air cooling.

Liquid cooling for CPULower CPU temperature and quieter sustained workstation use.+$0Air cooling for CPUSimpler service profile for moderate sustained loads.+$0

Memory

Size system RAM around the model, dataset, and total GPU VRAM.

64 GB DDR5Practical baseline for one GPU and common local model workflows.+$096 GB DDR5More room for notebooks, retrieval pipelines, and larger context jobs.+$337128 GB DDR5Preferred if GPU VRAM grows or datasets live in memory.+$648192 GB DDR5Maximum local-memory profile for this compact platform.+$1,663

GPU Support

Select the power and motherboard plan before choosing final GPUs.

Ready for 1x RTX 5070Lower-power single-GPU development profile.+$0Ready for 1x RTX 5090 or RTX PROHigher-power single-card local LLM profile.+$553Ready for 2x RTX 5090 or RTX PRODual-GPU expansion path with upgraded board and power budget.+$1,843

Storage

Keep the OS, model weights, checkpoints, and datasets on fast local storage.

2 TB PCIe 4.0 NVMe SSDFast local baseline for OS, tools, and model weights.+$04 TB PCIe 4.0 NVMe SSDMore room for datasets, vector stores, and checkpoints.+$4248 TB PCIe 4.0 NVMe SSDLarge local scratch and project storage.+$1,502

Operating System

Choose the installed OS image or keep the system ready for your own deployment process.

Ubuntu AI stackLinux image prepared for CUDA, containers, notebooks, and model-serving tools.+$0Windows 11 ProWindows desktop environment for mixed AI, engineering, and creative tools.+$376No operating systemFor teams that image workstations internally.+$0

Additional SSD

Add a separate data volume for datasets, checkpoints, and project files.

Not includedUse only the primary NVMe drive.+$04 TB SATA SSDSeparate project storage for less speed-sensitive files.+$1,46215.36 TB PCIe NVMe SSDLarge high-speed local dataset and checkpoint drive.+$13,176

Network

Plan bandwidth for shared datasets, remote access, and lab storage.

Integrated EthernetUse the platform's base network interface.+$010 Gb EthernetBetter fit for NAS access and shared lab environments.+$34425 Gb EthernetHigher-throughput network path for larger datasets.+$592

Warranty/Support

Choose the service window for the system lifecycle.

3 year limited warrantyBase service plan for the configured workstation.+$04 year limited warrantyLonger labor and replacement coverage.+$7515 year limited warrantyLongest support window for teams standardizing on one platform.+$1,145

Why this configuration path

Local LLM workstation profile with up to 192 GB VRAM.

Compact tower format with liquid-cooled CPU options.

Good entry point for private inference and model experiments.