The Novatech Apex WS9995X is the pinnacle of workstation performance, engineered for enterprise professionals, researchers, and creators who demand uncompromising power. At its heart is the AMD Ryzen Threadripper PRO 9995WX, a 96-core, 192-thread CPU designed to tackle the heaviest workloads with unmatched efficiency.
Paired with the NVIDIA RTX PRO 6000 GPU with 96GB of ECC VRAM, the WS9995X delivers exceptional acceleration for AI training, machine learning, HPC, 3D rendering, engineering simulations, and scientific computing.
With a massive 512GB of DDR5 ECC memory and 10TB of blazing-fast NVMe Gen 5 SSD storage (2TB + 4TB + 4TB), this system ensures seamless multitasking, rapid data access, and the capacity to handle the largest datasets and projects.
Cooling is managed by an ASUS ProArt 420mm liquid cooling system, while a 1200W 80+ Gold PSU provides stable, efficient power delivery. The full-tower SilverStone SETA H2 chassis offers professional expandability and airflow for future growth.
Every Novatech workstation is assembled and stress-tested in the USA, backed by lifetime technical support and a 3-year limited hardware warranty.
Whether you’re driving AI innovation, running advanced simulations, producing 8K video, or visualizing complex 3D models, the Apex WS9995X stands as the ultimate platform for professionals who refuse to compromise.
[Extreme AI & Machine Learning Performance]
Built with AMD Ryzen Threadripper PRO CPUs (up to 96 cores) and NVIDIA RTX GPUs including the RTX PRO 6000, RTX 5090, and RTX 5080. Designed for deep learning, neural networks, and AI model training with CUDA acceleration.
[Scalable Memory & Storage]
Choose configurations with up to 512GB DDR5 ECC RAM and 10TB of ultra-fast NVMe Gen 5 SSD storage. Perfect for data science, predictive analytics, and large-scale simulations requiring maximum speed and stability.
[Professional 3D Rendering & Design]
Handle demanding CAD, architecture, 3D modeling, and real-time rendering with workstation-class GPU power. Optimized for creative professionals and design studios requiring high-fidelity visuals and reliable compute.
[Gaming & Content Creation Powerhouse]
Play the latest AAA titles in 4K with ultra settings while seamlessly handling 8K video editing, streaming, and digital content creation. Balanced for both professional workloads and elite gaming performance.
[Buy With Confidence – Assembled & Supported in the USA]
All Novatech workstations are assembled and stress-tested in the USA, backed by lifetime technical support and a 3-year limited hardware warranty for guaranteed peace of mind.
Couldn't load pickup availability
Enterprise-aligned estimates for model fit, inference throughput, diffusion performance, and cloud ROI framing.
Capability
| Model / Class | Quantization | Inference | Fine-tuning |
|---|---|---|---|
| Llama 3 70B | FP16 / 4-bit | ✅ | ✅ |
| Mixtral 8x22B | FP16 / 4-bit | ✅ | ✅ |
| 13B–34B class | FP16 | 🚀 | ✅ |
| 405B (sharded) | 4-bit | ✅ | ⚠️ |
Legend
Actual capability varies by framework, context length, batch size, and optimization stack.
Throughput
| Model | Precision | Estimated tok/sec |
|---|---|---|
| 7B | FP16 | 600–900 tok/sec |
| 13B | FP16 | 350–600 tok/sec |
| 34B | FP16 | 180–300 tok/sec |
| 70B | FP16 | 120–200 tok/sec |
| 70B | 4-bit | 220–350 tok/sec |
Estimates assume an optimized inference stack (e.g., vLLM / TensorRT-LLM). Batch inference can materially increase throughput.
Generative
| Workflow / Model | Estimated performance |
|---|---|
| Stable Diffusion XL | ~1–2 sec/image (optimized) |
| SD Turbo | Sub-second generations (pipeline dependent) |
| Video diffusion (optimized) | Near real-time depending on resolution |
Render time depends on resolution, steps, scheduler, and pipeline optimizations.
ROI
| Scenario | Estimate |
|---|---|
| H100-class cloud instance | $90–$110/hr (varies by region/commit) |
| 8 hrs/day utilization | ~$20k–$26k/month |
| Directional break-even | Often ~4–6 months (utilization dependent) |
Cloud rates vary by region, commitment, and availability. Use as directional procurement framing.
Deploy an on-prem AI server or enterprise AI workstation designed for LLM training hardware, fine-tuning, and secure internal workflows—without giving up performance or control.
Max configuration highlight
Designed for teams that want predictable throughput and private data workflows—scalable from a strong base config to a fully-loaded build.
Base configurations start smaller (e.g., 1× GPU / 128GB RAM / 1TB), then scale to match your model size, context, and concurrency requirements.
Use cases
Designed for teams that need private AI infrastructure—where governance, predictable throughput, and dedicated GPU capacity matter.
Host a private LLM for engineering, support, HR, or ops—behind your firewall.
Iterate on domain-specific models with controlled datasets and repeatable benchmarking.
Run GPU-heavy workflows for QA, inspection, security review, and media analytics.
Serve multiple users or multiple models with headroom for peak usage.
Why on-prem?
If workloads are steady—or data can’t leave your environment—on-prem can be the simplest path to consistent performance and privacy.
Need hybrid? Many teams keep sensitive retrieval and inference on-prem while bursting less sensitive training jobs to cloud.
Important
We’ll tell you if you’re better served by cloud or a smaller system. This is the right fit when you need dedicated GPU infrastructure and plan to use it.
Contact
Get a quick recommendation and a quote based on your model, throughput needs, and on-prem requirements.
Ready for sizing & pricing?
Email us your model, expected users/concurrency, and whether data must stay on-prem.
Fast path: include your model name, expected users/concurrency, and whether your data must remain on-prem.
Directional, production-aligned expectations for 3D rendering, simulation, and professional video workflows. Customize the tables per product configuration.
Works great with
3D Rendering
| Software | Workload | Expected performance class |
|---|---|---|
| Blender (Cycles) | GPU path tracing / final-frame rendering | ✅ Designed to handle complex scenes; performance varies by scene fit and settings |
| Cinema 4D + Redshift | GPU rendering + lookdev | ✅ Strong RTX throughput; scaling and results depend on scene/renderer |
| OctaneRender | Photoreal path tracing | ✅ Built for heavy RTX workloads; multi-GPU behavior depends on your scenes |
| Unreal Engine 5 | Nanite + Lumen viewport | ✅ Designed for large projects; GPU + RAM help complex assets and editor responsiveness |
Use these as directional expectations. Actual performance varies by renderer settings, scene fit in VRAM, and effects complexity.
Legend
Video Production
| Software | Workflow | Typical workflows supported |
|---|---|---|
| DaVinci Resolve | 4K–8K editing + grading (GPU accelerated) | ✅ Built for GPU-accelerated editing and effects; codec/effects dependent |
| Adobe Premiere Pro | Multi-stream 4K editing + GPU effects | ✅ Excellent workflow support; depends on codecs and effects stack |
| After Effects | Large comps + RAM-heavy projects | ✅ Designed to handle large comps with heavy caching (project dependent) |
Results depend on codec, effects stack, GPU acceleration settings, and storage throughput.
Large Project Handling
| Resource | Benefit |
|---|---|
| 192GB RAM | Massive scenes, large After Effects comps |
| Up to 10TB NVMe storage | Fast project and asset loading |
| RTX 5090 / RTX Pro 6000 VRAM | Large textures, heavy scenes |
| Ryzen 9950X3D | High viewport and CPU simulation performance |
Designed to handle large textures, heavy scenes, and RAM-intensive comps—especially when projects stay on fast NVMe storage.
Render Speed Comparison
| System | Relative render speed |
|---|---|
| Typical RTX 5080 workstation | 1× |
| Single RTX 5090 or RTX Pro 6000 workstation | ~1.2–1.4× |
| Dual high-end GPU workstation (same class) | ~2.0–2.6× |
Directional comparison for GPU-rendered workloads. Actual results vary by renderer support, scene fit, and multi-GPU scaling behavior.
Note: All guidance is directional. Actual results vary by scene complexity, codecs, effects, renderer settings, and software versions.
Cut model training times and speed up simulations with the AMD Threadripper PRO 9995WX (96 cores, 192 threads) and NVIDIA RTX PRO 6000 with 96GB VRAM.
Run AI, HPC, data analytics, 3D rendering, and engineering simulations without bottlenecks — keeping teams productive and projects on schedule.
With 512GB DDR5 ECC memory, 10TB of Gen 5 NVMe storage, and a future-ready WRX90E platform, your workstation grows with your company’s needs.
Is it powerful enough for machine learning?
Can it handle big datasets and multitasking?
Is it reliable for long workloads and 24/7 operation?
Will it stay relevant as technology advances?
Is this workstation only for AI?
