A local AI rig shows the real cost of running models yourself
A high-end personal computer for running locally is now complete. The original plan was to buy an in the United States for about $8,500, with a possible $1,000 discount through the NVIDIA Inception program. The took about three months and was rejected, and during that delay the card’s price rose to $13,500, which put it out of budget.
The final choice was an RTX Pro 5000 found locally as the last available unit. The finished machine has 80GB of VRAM, 192GB of RAM, 17TB of storage, a 9950X3D processor, and a 1.3kW . Running both the and processor at full power could use about 95% of the available power, so limiting the RTX 5090’s power may be needed for safety.
The machine is intended for Q_8s models and work.
Key points
- The planned rose from about $8,500 to $13,500 during a long discount-program wait.
- The NVIDIA Inception discount was rejected after about three months.
- The final build used a locally available RTX Pro 5000 instead.
- The completed machine has 80GB of VRAM, 192GB of RAM, and 17TB of storage.
- Power use could get close to the limit, so power limiting may be needed.