xAI Colossus supercomputer (Memphis, Tennessee)

xAI’s Colossus supercomputer is in Memphis, Tennessee.

Owner
xAI
Partners
MLGW and TVA (utilities)
Location
Memphis, Tennessee
Capacity as announced
xAI states 200,000 GPUs; no power capacity stated
Chips
NVIDIA H100 and other NVIDIA GPUs
Status
Operational
Timeline
Built in 122 days, then doubled in 92 days; xAI states a roadmap to 1 million GPUs

xAI’s Colossus page says the system reached 200,000 GPUs, was built in 122 days and then doubled in 92 days, with 170 PB/s of total memory bandwidth, 2.8 Tb/s of network bandwidth per server and more than 0.5 EB of storage, and it describes a roadmap to 1 million GPUs. xAI’s Memphis page names MLGW and TVA as its utilities. Neither page states a power capacity. Our explainer: xAI Colossus.

Country
United States
Project type
Data center campus
Announcement
x.ai
Checked

Questions

How many GPUs does xAI Colossus have?
xAI states 200,000 GPUs and a roadmap to 1 million.
How long did Colossus take to build?
xAI says it was built in 122 days and doubled in another 92 days.
Where is Colossus?
In Memphis, Tennessee, served by MLGW and TVA.

Source

Every value comes from the company’s own announcement, investor release or newsroom post, linked on each project with the date it was last checked. How we pick and check sources: editorial standards.

The Model Press

What are you looking for?

Search by headline, topic or keyword.