Desktop GPU

NVIDIA RTX 5090

The Fastest Consumer GPU Ever Made

~$4,500 (street price, up from $1,999 MSRP due to GDDR7 shortage)
Market estimate
32 GB
GDDR7 VRAM
32 gigabytes of fast video memory. That's the workspace the GPU uses to hold model weights while generating tokens. Bigger = can run larger models without swapping to slow regular RAM. At full precision, this fits models up to about 25 billion parameters.
575 W
Total Power Draw
Thermal Design Power — how much electricity it draws under load. 575 W is roughly the same as running four gaming consoles at once, or about half an average home's continuous draw (which is ~1,200 W).
~$4,500 (street price, up from $1,999 MSRP due to GDDR7 shortage)
Street Price
MSRP at launch (January 2025) was $1,999, but a GDDR7 memory shortage drove street prices to $4,300–$4,900 by August 2026. You're paying a supply-chain premium — if that gap concerns you, consider waiting or stepping up to the RTX PRO 6000 Blackwell for more memory at a more predictable enterprise price.

⚡ Power in everyday terms

🏠 0.48× average homes running simultaneously (avg home = 1,200 W continuously)
🔋 Drains a 90 kWh EV battery in 156.5 hours

✅ Good for

  • Inference on models up to 25 B params
  • Small teams experimenting with open-source LLMs
  • Single-user workstations

❌ Not the right fit for

  • The three models on our AI Model page (all require 340+ GB)
  • Any production workload needing high throughput
Form factor: PCIe Card (fits in a desktop)

Also consider