NVIDIA B200

NVIDIA B200

NVIDIA's current-generation datacenter AI chip. Sold inside servers, not by itself.

GPU memory

180 GB

GPU memory is the "workspace" a graphics card uses to hold an AI model while it runs. If a model is bigger than the available memory, it simply will not fit — like trying to fit a large book on a shelf that's too small. More in the wiki →

Power draw

1,000 W

Watts measure how much electricity something draws while it's running. More powerful hardware needs more electricity and produces more heat that has to be cooled. More in the wiki →

Ideal cooling

Passive - server airflow (liquid variants exist)

Cooled by the host server's fans; the densest deployments use liquid-cooled variants instead. Running flat out, it turns 1,000 W of electricity into heat that has to go somewhere. We stock matching cooling gear →

Price

$35,000

This is a current typical market price, not a fixed list price — real prices move with supply and demand. NVIDIA CEO Jensen Huang stated on CNBC that Blackwell GPUs would cost $30,000–$40,000 each. We use the middle of that range as a per-GPU price.

What this means in everyday terms

Running one of these draws as much power as about 0.8 average homes running around the clock. A full day of running one uses about 0.27 electric-car batteries worth of energy. An average home draws roughly 1,200 watts around the clock. This compares a build's power draw to that baseline. More in the wiki →

NVIDIA B200 is sold to run inside a server or rack — it isn't something you plug into a home PC.

Where these numbers come from

Price per Jensen Huang's on-record CNBC statement. Memory/power per Lenovo Press ThinkSystem NVIDIA B200 spec sheet.

https://www.cnbc.com/2024/03/19/nvidias-blackwell-ai-chip-will-cost-more-than-30000-ceo-says.html