Skip to main content
  • $30,999.00
  • $32,999.00
  • Save: $2,000.00 (6%)
  • Sold by Newegg

    Shipped by Newegg

    Price alert
    Shop it at NeweggBusiness.

    ABS AI Workstation - 96GB VRAM (NVIDIA RTX PRO 6000 Blackwell), Intel Xeon W9-3575X, 128GB ECC DDR5, 2TB NVMe M.2 + 3.84TB SATA SSD, Ubuntu (Zaurion ZAW9-3575X-RP6000)

    • Threadripper 9965WX | 3x Arc Pro B70 | 128GB DDR5 | 2x 2TB M.2
    • Threadripper 9965WX | 3x Radeon AI Pro R9700 | 128GB DDR5 | 2x 2TB M.2
    • Threadripper PRO 7975WX | RTX Pro 6000 Blackwell | 128GB DDR5 | 2TB+4TB SSD
    • Threadripper PRO 7975WX | RTX Pro 6000 Blackwell | 64GB DDR5 | 1TB+2TB SSD
    • Threadripper PRO 7975WX |2x RTX Pro 6000 Blackwell MaxQ | 128GB DDR5 | 2TB+4TB SSD
    • Xeon W5-2455X | RTX Pro 6000 Blackwell | 64GB DDR5 | 1TB+2TB SSD
    • Xeon W9-3575X | 2x RTX Pro 6000 Blackwell MaxQ | 128GB DDR5 | 2TB+4TB SSD
    • Xeon W9-3575X | RTX Pro 6000 Blackwell | 128GB DDR5 | 2TB+4TB SSD
    • Compliance Documents Required for Purchase
    • Assembled in USA
    • TAA Compliant
    • Operating System: Ubuntu
    • CPU: Intel Xeon W9-3575X
    • GPU: RTX Pro 6000 Blackwell (96 GB GDDR7)
    • Motherboard: Gigabyte MW83-RP0
    • Memory: 128GB DDR5 Server RAM - ECC, Registered
    • OS SSD: 2TB M.2 NVMe (Installed)
    • Data SSD: 3.84TB SSD (Installed)
    • LAN: 2 x 10GB/s LAN, 1 x 10/100/1000 Mbps Management LAN
    • PSU: Single 2000W ATX 80 PLUS Gold power supply
    • Barebone: Gigabyte W773-W80
    • Optional: Pascari aiDAPTIV SSD (for VRAM expansion & large model fine-tuning)
    • AI Agent Ready: Optional pre-installed OpenClaw or NemoClaw (enterprise AI agent framework)
    • Local LLM Support: Optional Qwen 3.5 27B or higher models for on-prem AI inference
    This item is covered under the Manufacturer Only Return Policy.
    +
    +
    Overview
    Specs
    Reviews
    See all Q & A
    avatar

    Any questions? Our AI mode will help you find out quickly.

    ABS Zaurion Aqua · AI Workstation · ZAW9-3575X-RP6000

    Still Waters. 4000 TOPS Deep.1,2

    Intel Xeon W9-3575X  ·  NVIDIA RTX PRO 6000 Blackwell  ·  96 GB GDDR7

    Stop renting GPUs. The Zaurion Aqua puts data-center silicon in a quiet tower: 96 GB of GDDR7 on a single professional Blackwell card, a flagship Xeon W backbone, and Ubuntu ready for CUDA on first boot. Your models, your data, your hardware — zero cloud bills.

    AI Ready Runs Large Local LLMs Ubuntu Pre-Installed Assembled in USA TAA Compliant
    Illustrative scene — a quiet tower workstation serving a private local AI assistant at night
    96 GB
    GDDR7 GPU memory
    128 GB
    DDR5 ECC Registered
    5.84 TB
    SSD storage installed
    2×10 GbE
    + management LAN
    2000 W
    80 PLUS Gold PSU
    Module 02 · Explore the Machine

    Every Port Has a Job. Hover the Dots.

    The rear panel reads like a datacenter spec sheet. Hover each marker (or read the legend) to see what this workstation-class I/O actually buys you.

    Rear I/O of the Zaurion Aqua — networking, quad DisplayPort, PCIe expansion, and 2000 W PSU 1 Dual 10 GbE + management LANWire-speed dataset transfers to your NAS or cluster, plus out-of-band IT control. 2 Directed exhaust airflowEngineered cooling keeps CPU and GPU at full boost through multi-day runs. 3 Quad DisplayPort on the RTX PRO 6000Drive multi-monitor and 8K-class displays with DisplayPort 2.1. 4 Full-height PCIe expansionServer-class Gigabyte MW83-RP0 board with room for future accelerators and NICs. 5 2000 W 80 PLUS Gold PSUEfficient, stable headroom for sustained full-load GPU training.
    Rear panel, actual product photo — hotspot positions are content-editable.
    • 1
      Dual 10 Gb/s LAN + management port

      A 100 GB corpus lands in minutes, not hours — and IT can power-cycle and monitor the box remotely, even when the OS is down.

    • 2
      Directed exhaust airflow

      Internal airflow diversion sustains full boost clocks around the clock.

    • 3
      Quad DisplayPort (RTX PRO 6000 Blackwell)

      DisplayPort 2.1 bandwidth for color-accurate multi-monitor and 8K-class work.

    • 4
      Full-height PCIe expansion

      The server-grade Gigabyte MW83-RP0 platform leaves lanes for tomorrow's cards.

    • 5
      2000 W 80 PLUS Gold power

      One PSU, zero drama — headroom for the GPU, CPU, and everything you add.

    Module 03 · Usage Scenarios

    Six Jobs It Does on Day One

    Business and prosumer workloads this exact configuration was built for — no hypotheticals.

    Illustrative scene — AI startup office with deskside tower workstation and LLM dashboards

    Private ChatGPT-Style Assistant

    Serve large open-weight models to the whole team with an OpenAI-compatible API. Zero per-token bills; nothing leaves your firewall.

    Qwen / Llama on vLLM or Ollama · long-context chat for dozens of users
    Illustrative scene — research lab with training-loss and GPU-memory charts on a wall display

    Fine-Tune on Your Own Data

    Adapt models to your domain — support transcripts, contracts, catalogs. LoRA fits in 96 GB; optional aiDAPTIV unlocks full fine-tuning of ~34B models.3

    LoRA / QLoRA in PyTorch · ECC keeps multi-day runs clean
    Illustrative scene — analyst querying an enterprise knowledge base with cited document results

    RAG Knowledge Base

    Index wikis, tickets, and manuals; employees ask in plain language, answers come back grounded with sources — all on-prem.

    Embeddings + reranking on-GPU · 3.84 TB data SSD for the corpus
    Illustrative scene — operations engineer monitoring an AI agent pipeline dashboard with auto-triaged tickets

    Agentic AI Automation

    Agents that read tickets, draft replies, and chain multi-step tasks — optionally shipped with OpenClaw or NemoClaw pre-installed.

    Support triage · code-review bots · MIG isolates each agent
    Illustrative scene — engineer with a ray-traced CAD model and physics simulation heat map

    Rendering & Simulation

    4th-gen RT Cores drive photoreal ray tracing for product design, AECO, and digital twins; 96 GB holds scenes smaller cards can't open.

    Blender / V-Ray / Omniverse · 8K-class output via DisplayPort 2.1
    Illustrative scene — color-grading and editing studio with dual monitors and tower workstation

    Pro Video & Streaming

    9th-gen NVENC and 6th-gen NVDEC accelerate encode and ingest, including 4:2:2 H.264/HEVC, for editing suites and live production.

    Multi-stream 4K/8K editing · AI upscaling & enhancement

    Scene photos are AI-generated illustrations of typical deployments, not photographs of this exact product.

    Module 04 · Comparison Highlights

    The Numbers That Matter

    4000 TOPS
    AI Performance1,2
    380 TFLOPS
    RT Core Performance1
    125 TFLOPS
    Single-Precision1
    This workstation — RTX PRO 6000 Blackwell
    96 GB
    Typical single-GPU workstation
    24 GB

    Illustrative comparison of usable GPU memory for model loading.

    Storage ships ready for real work — a 2 TB M.2 NVMe system drive plus a 3.84 TB data SSD — and the optional Pascari aiDAPTIV SSD tier rewrites the fine-tuning math:

    Target Model SizeVRAM Needed (Without aiDAPTIV)VRAM Needed (With aiDAPTIV)
    ~13B120–200 GB32 GB
    ~34B250–350 GB96 GB — this workstation's GPU
    ~70B500–700 GB192 GB

    Pascari aiDAPTIV SSD is an optional add-on and is not included by default. VRAM figures are vendor-published estimates for full fine-tuning.3

    Module 05 · Customer Use Cases & Options

    Deploy It Your Way

    Illustrative deployment profiles — and the Build-to-Order options that fit each one:

    Illustrative scene — IT administrator cabling the rear panel with a management console laptop
    Rolls in like a desktop, managed like a server — out-of-band control via the dedicated management LAN. Illustrative scene.
    Profile

    The AI Product Team

    Private LLM serving + fine-tuning on customer data

    Inference by day, LoRA runs by night. Pair with the Local LLM Package (Qwen 3.5 27B or higher, pre-configured) and skip setup week entirely.

    Profile

    The Research Lab

    Shared GPU for a team, procurement-friendly

    MIG gives each researcher an isolated GPU slice; TAA compliance and USA assembly clear public-sector purchasing. Add the Pascari aiDAPTIV SSD for full fine-tuning headroom.3

    Profile

    The Automation Builder

    Agentic AI in production

    Order it with OpenClaw or NemoClaw pre-installed and ship ticket-triage and code-review agents in week one, isolated per MIG partition.

    Sibling Config

    Step Up: Dual GPU

    Same Xeon W9-3575X platform with 2× RTX PRO 6000 Blackwell MaxQ for maximum combined GPU memory — compare options on this listing.

    Sibling Config

    Value Entry: Xeon W5

    Xeon W5-2455X with the same 96 GB Blackwell GPU — the budget door into large-model territory.

    Build to Order

    Custom Built After Order

    Every unit is assembled and stress-tested in the USA to your configuration — request a quote or customize from this page.

    Module 06 · FAQ

    Quick Answers Before You Buy

    Can it really run large language models locally?

    Yes. 96 GB of GDDR7 on a single card loads large open-weight models (Qwen, Llama, and similar) for inference and LoRA fine-tuning; the optional aiDAPTIV tier extends full fine-tuning to ~34B-class models.3

    What operating system does it ship with?

    Ubuntu comes pre-installed — the native home of CUDA, PyTorch, vLLM, and the open-source AI stack.

    Can several people share one machine?

    Yes. Universal MIG partitions the GPU into isolated instances, and 2 × 10 Gb/s LAN ports serve model APIs to your whole team at wire speed.

    Is it eligible for government / public-sector purchasing?

    It is TAA compliant and assembled in the USA, with compliance documents available for purchase review.

    How is it supported?

    1-year parts and labor limited warranty backed by ABS manufacturer support. As a Build to Order system, it is custom built and stress-tested after purchase.

    Assembled in USA
    TAA Compliant
    Server-grade ECC platform
    Build to Order flexibility
    1. Peak rates based on GPU Boost Clock.   2. Theoretical FP4 TOPS using sparsity. NVIDIA performance figures as published by NVIDIA for the RTX PRO 6000 Blackwell.   3. VRAM requirement figures are vendor-published estimates for full fine-tuning with the optional aiDAPTIV tier; actual results vary by model and configuration. ABS Zaurion Aqua ZAW9-3575X-RP6000 · Optional components and software are configured at time of purchase.

    Warranty & Returns

    Warranty, Returns, And Additional Information

    Warranty

    • Limited Warranty period (parts): 1 year
    • Limited Warranty period (labor): 1 year
    • Read full details

    Return Policies

    LOADING...