Skip to content

The local AI publication

Workstation build · mid tower · New parts

Professional AI Workstation

High-end single-GPU workstation

Built by Llama Builds · Updated

White case interior with a dual-fan NVIDIA reference-style graphics card, an AIO pump with an LCD temperature display, RGB fans, and rainbow-lit power cable extensions.
Build estimate
$10,397
VRAM
48GB
GPUs
1
System RAM
64GB
View the parts list

Build prices are historical estimates. Follow each retailer link for current pricing.

What can this build run?

With 48GB of VRAM across one GPU, this build comfortably runs models up to Qwen3.6 35B-A3B (Q8_0, ~100 tokens/sec). Our top all-round pick for it is Qwen3.8 27B at Q8_0 (~25 tokens/sec).

Parts list

Every part in the Professional AI Workstation

Prices at time of build (2025); check current prices.

We may earn a commission when you buy through links on this page. It never changes what we recommend. How we make money

Parts, historical estimates, and retailer links for Professional AI Workstation
PartBuild estimateCurrent availability
GPUNVIDIA RTX 6000 Ada GenerationBuild estimate$8,999Search Amazon
CPUAMD Ryzen 9 7900XBuild estimate$329View at Amazon
MotherboardNZXT N7 B650EBuild estimate$190Search Amazon
Power SupplyCORSAIR RM1000x 80+ GoldBuild estimate$180View at Amazon
Memory2×Corsair Vengeance DDR5-5600 32GBBuild estimate$240$120 eachView at Amazon
StorageSamsung 980 PRO 1TB NVMe SSDBuild estimate$130Search Amazon
CaseNZXT H7 Flow Mid-TowerBuild estimate$129View at Amazon
CPU CoolerNZXT Kraken 280 RGB AIOBuild estimate$200View at Amazon
Estimated build total$10,397

Capability

Models this build runs well

Full list

Estimated for a single user with 8K context. Best quantization that fits, ranked by our capability score.

  • Best overall · also coding, reasoning, vision

    Qwen3.8 27B

    Q8_0 · ~25 tokens/sec

  • Runner-up · also largest that fits

    Qwen3.6 35B-A3B

    Q8_0 · ~100 tokens/sec

  • Fastest good model

    gpt-oss-20b

    MXFP4 · ~130 tokens/sec

ModelQuantSpeedFit
Qwen3.8 27BQ8_0~25 tok/sRuns well
Qwen3.6 35B-A3BQ8_0~100 tok/sRuns well
Gemma 4 31BQ8_0~23 tok/sRuns well
Gemma 4 26B A4BQ8_0~89 tok/sRuns well
gpt-oss-20bMXFP4~130 tok/sRuns well
Qwen3.5 9BQ8_0~66 tok/sRuns well
Gemma 4 E4BQ8_0~75 tok/sRuns well
Phi-4Q8_0~47 tok/sRuns well

The build

About this build

A clean single-GPU workstation built around the 48GB NVIDIA RTX 6000 Ada Generation for enterprise AI development and training, on a mainstream Ryzen 9 platform.

Why this build

A clean, professional build featuring the NVIDIA RTX 6000 Ada Generation for enterprise AI development and training workloads. The card’s 48GB of VRAM does the heavy lifting, while a Ryzen 9 7900X, 64GB of DDR5, and an NZXT H7 Flow case keep the rest of the system straightforward.

Specifications

Full specs

GPU
RTX 6000 Ada
CPU
AMD Ryzen 9 7900X
Motherboard
NZXT N7 B650E
Memory
64GB DDR5-5600 (2 x 32GB Corsair Vengeance)
Storage
Samsung 980 Pro 1TB NVMe SSD
Power supply
Corsair RM1000x 80+ Gold
Case
NZXT H7 Flow (mid tower)
Cooling
NZXT Kraken 280 RGB AIO
Form factor
mid tower
GPU power
300W rated

Keep browsing

Similar builds

All builds
Gold-finished workstation case with perforated panels, a glass side panel, an A16Z logo, and caster wheels, displayed in an office lobby with plants and a backlit sign.
$43,568

Workstation · 384GB VRAM

a16z Personal AI Workstation

Cutting-edge 4x RTX Pro 6000 Blackwell workstation

4× RTX PRO 6000

Open-air AI rig built around a small white IKEA table, with MSI Suprim RTX 3090 cards on top and on the floor, the motherboard underneath, and several power supplies with thick cable bundles beside it.
$6,486

High-end · 96GB VRAM

FJÄLLBO

AI workstation arranged on an IKEA table

4× RTX 3090

Frame from PewDiePie's video showing his black aluminum-extrusion open rack with three large top fans, dense power cabling, and Seasonic Prime power supplies.
$20,429

Server · 140GB VRAM

PewDiePie "Odysseus"

PewDiePie's first NVIDIA multi-GPU inference rig

7× RTX 4000 Ada