Autonomous Computer Rack

From

$93,900

Autonomous Computer Rack

On-prem AI for the company. 8 GPUs.

BuildBuild: Complete machine · 8× RTX 5090

Complete machine · 8× RTX 5090

VRAM · 256 GB

FP32 · 838 TFLOPS

Case · 7U rack, hot-swap

Risers · 8× MCIO Gen5 ×16

Board · ASRock TURIN2D24G-2L+

CPU · 2× AMD EPYC 9124

Power · 2000W CRPS, 240V

Cooling · 12 hot-swap fans

Complete machine · 8× RTX PRO 6000

DIY kit · Case and PCIe 5.0 risers

Memory

256 GB

512 GB

768 GB

Storage

1 TB

2 TB

4 TB

8 TB

Buy more, save more

Qty	Price
3-4	$91,083
5-9	$89,205
10-29	$88,266
30+	$87,327

$93,900

Free shipping within 2-4 weeks

Autonomous Computer Rack

Free shipping within 2-4 weeks

$93,900

Own your intelligence.

Never revoked.

Your models don't get deprecated, rate-limited, or pulled. The weights sit on your disk.

Private by physics.

Prompts never leave the building. Nobody trains on your work. Not by policy, by physics.

No per-token bill.

Own the machine and agents can run around the clock. You pay for electricity, not tokens.

Yours to keep.

Fine-tune on what only you know. The model becomes an asset you own, and it sells with the company.

GPUs

Eight cards, tuned and burn-tested. We install RTX 5090 or RTX PRO 6000 Blackwell — installed, driver-loaded, and benchmarked as one system before it ships. Plug into a 240V circuit and run.

256-768GB

Total VRAM

838–1006TFLOPS

FP32

14,336GB/s

Memory bandwidth

Chassis

A server chassis, not a showpiece. At eight cards you're building something that lives in a rack and runs for weeks, so it ships in a 7U GPU server chassis built for exactly this: hot-swap fans, hot-swap drive bays, pull a drive without opening the machine.

Rack height

Hot-swap NVMe bays

93lbs

Weight

Riser

Datacenter cabling, not desktop ribbon. Each GPU connects over a PCIe 5.0 ×16 riser fed by two MCIO cables — the shielded connector datacenters use to carry Gen5 without dropping signal. Straight from the CPU, no PCIe switch deciding which card waits.

x16

Gen5 lanes per card

2xMCIO cables per card

Per card

PCIe switches

Motherboard

ASRock Rack TURIN2D24G-2L+. A dual-socket AMD EPYC server board with enough Gen5 lanes to feed all eight cards direct from the processors — plus server-grade remote management, so you can watch it and restart it without standing in front of the rack.

SP5 sockets

DIMM slots

BMC

Remote management

CPU

Two AMD EPYC 9124s. Dual server processors carry every card on its own dedicated Gen5 ×16 — no sharing, no waiting. This is the same silicon that runs the datacenters you're replacing.

EPYC 9124

Cores total

Gen 5

Direct to every card

Power

Built to stay up. 2000W CRPS server supplies, hot-swap and redundant — pull a failed unit while the box keeps serving. One thing to plan for: this needs a 240V circuit, not a wall outlet. Sort it before the machine arrives.

2000W

Per CRPS unit

Hot-swap

Redundant supplies

240V

Circuit required

Cooling

Air moved on purpose. A wall of twelve hot-swap fans, PWM controlled, wired to one controller and matched to the airflow path through the case. Swap a fan without stopping a job.

Memory & Storage

ECC everywhere, 8TB of NVMe. Sixteen 32GB DDR5-4800 ECC RDIMMs catch silent bit-flips on jobs that run untouched for weeks. Two 4TB NVMe drives in hot-swap bays, six bays free to grow.

512GB

ECC RAM

8TB

NVMe, hot-swap

Bays free

GPU

Cards

8x NVIDIA RTX 5090 or 8x NVIDIA RTX PRO 6000 Blackwell

Total VRAM

256 GB (RTX 5090) / 768 GB (RTX PRO 6000)

FP32 compute

838 TFLOPS (RTX 5090) / 1006 TFLOPS (RTX PRO 6000)

Interconnect

PCIe 5.0 x16 per card via 2x MCIO cables, direct to CPU - no PCIe switch

Compute

CPU

2x AMD EPYC 9124 - 32 cores total

Motherboard

ASRock Rack TURIN2D24G-2L+ - dual SP5, 24 DIMM slots

Remote management

BMC, server-grade

CPU cooler

2x COOLSERVER 2U SP5 S22

Memory & storage

RAM

512 GB (16x 32GB DDR5-4800 ECC) - configurable 256-768 GB

Storage

2TB NVMe, hot-swap bay - configurable 1-8TB

Drive bays

8 hot-swap NVMe bays, 6 free

Power & cooling

Power supply

2000W CRPS, hot-swap redundant

Wall power

240V circuit required

Fans

12x hot-swap, PWM controlled

Dimensions

Size

17.3"W x 35.5"D x 12.2"H - 7U rack

Weight

93 lbs

Chassis

7U GPU server chassis, steel, hot-swap serviceable

Software

Ubuntu, NVIDIA drivers pre-installed

AI frameworks

PyTorch, CUDA, vLLM, SGLang, llama.cpp

Warranty

Full system

1 year

Other

Open source

CAD, BOM & BIOS on GitHub

Shipping

2-4 weeks

The lineup

Three builds.

Autonomous Computer (2-GPU)

12.5” × 12.5” × 16” · 33lb

RTX Pro 6000 Blackwell or RTX 5090 x2
Up to 192GB VRAM · 128GB ECC RAM
Intel Xeon w5-3423 · 2TB NVMe
CNC-machined aluminum, black
PCIe 5.0 riser x2 · 6-fan cooling
2000W power · standard outlet

Autonomous Computer Studio (4-GPU)

15.5” × 15.5” × 16” · 66lb

RTX Pro 6000 Blackwell or RTX 5090 x4
Up to 384GB VRAM · 192GB ECC RAM
Intel Xeon w5-3423 · 2TB NVMe
CNC-machined aluminum, silver
PCIe 5.0 riser x4 · 18-fan cooling
2x 2000W power · 240V circuit

Autonomous Computer Rack (8-GPU)

17.3” × 35.5” × 12.2” · 93lb

RTX Pro 6000 Blackwell or RTX 5090 x8
Up to 768GB VRAM · 512GB ECC RAM
2x AMD EPYC 9124 · 8TB NVMe hot-swap
Steel 7U chassis, hot-swap everything
8x PCIe 5.0 x16 via MCIO · 12 hot-swap fans
2000W CRPS redundant · 240V circuit

Open-source hardware

The whole machine
is open.

Every CAD file, bill of materials, and BIOS setting — free on GitHub. Fork it, change it, build your own, even sell it. This is how the personal computer began: in the open. We’re doing it again, for AI hardware.

GitHub 1.2k

autonomous-ai / autonomous-computer

2x-5090/ 4x-5090/ 8x-5090/ 4x-6000/

bom/ step_models/ stl-models/ photos/

README.md setup.md

MIT · 147 forks

Clone it. Build it. Sell it. We just want it built.

Autonomous Computer Rack

On-prem AI for the company. 8 GPUs.