- 1 Cooling shroud + fans
- 2 GPU die + VRAM modules
- 3 PCIe interface + backplate
Desktop card
In stock
RTX 5090
A high-end desktop graphics card. The smallest, cheapest way to get real AI horsepower on your own desk.
SKU
GPU-RTX5090-32G
PriceiChecked against NVIDIA's own product listings and major retailers (B&H, Micro Center, NVIDIA's marketplace) as of July 2026 — real published prices, not estimates.
$1,999
Specifications
GPU memory
32 GB GDDR7iA fast memory type used on desktop and workstation graphics cards. Cheaper per gigabyte than the HBM used in datacenter GPUs, but with less total bandwidth.
Power draw
575 W
Tensor / AI compute
3,352 AI TOPS (FP4 sparse)
Rack units
N/A — desktop PCIe card, not rack-mounted
CUDA cores21,760
Memory bandwidth1,792 GB/s
ArchitectureBlackwell
Form factorDesktop PCIe card
Which models this can run
Real memory math against the 3 models in the model advisor — parameters × 1 GB, plus 20% working room.
In simple terms
Imagine an AI model's brain is made of billions of tiny sticky notes, and each one needs about
a gigabyte of shelf space. Every single sticky note has to fit on the computer's memory shelf
at the same time, with a little extra room to spare — if the shelf's too small, they just don't
fit, and the computer can't run that model at all, no matter how good the rest of it is.
Qwen3.5-397B
476.4 GB required · this unit has 32 GB
Does not have enough memory, and this product does not cluster.
DeepSeek-V3.2
805.2 GB required · this unit has 32 GB
Does not have enough memory, and this product does not cluster.
Kimi K2.6
1,200.0 GB required · this unit has 32 GB
Does not have enough memory, and this product does not cluster.
Power and cooling
Electrical requirement: Draws 575 W — fits within a single standard 120V/20A circuit (1,920 W safe continuous per NEC's 80% rule). No special electrical work needed.
About 0.5x the continuous draw of an average home
(1200 W). Run flat-out for a month, that's roughly
414 kWh — about $54/month
at a $0.13/kWh commercial rate (US average, EIA — your
actual rate depends on region and utility contract).
In simple terms
Watts are like how hungry a machine is for electricity — a bigger number means it "eats" more
electricity, kind of like a bigger animal eating more food. This one draws about as much
electricity as 0.5 whole houses running non-stop, which is
why bigger machines like this need special rooms built to feed and cool them.
Power cost, relative to the rest of the catalog
$54/mo
Cooling: Air-cooled (fan design varies by add-in card partner)
What this is good for
- Local inference and fine-tuning on models that fit in 32 GB
- Prototyping before committing to server-scale hardware
- A single-user development workstation
What this is NOT good for
- Training large models from scratch
- Serving many concurrent users at once
- Any model that needs more than roughly 26 GB after working-room overhead
Included components
GPU card only. No PSU, case, motherboard, or CPU included — this is a component for an existing system.
Lead time and warranty
Delivery window estimateTypically ships within 3–10 business days from authorized retailers, subject to stock.
WarrantyManufacturer warranty applies (terms vary by reseller).