Colossus 1

Colossus 1 is the supercomputer that trains Grok, the AI behind xAI's chatbot. It is one huge machine, housed in a roughly 785,000-square-foot former Electrolux appliance plant in Memphis, Tennessee, and it came online in September 2024 with 100,000 of NVIDIA's flagship H100 graphics chips. A graphics processing unit (GPU) is the specialized chip that does the heavy parallel math behind modern AI, and the whole thing was built, start to finish, in just 122 days, which is why Elon called it the most powerful AI training system in the world.

What is Colossus 1?

Colossus 1 is an AI training cluster, which means tens of thousands of GPUs wired together so they work as one enormous computer. It has one job, and that is to train the Grok family of large language models, the AI systems that read and write text inside xAI's assistant. Elon called it the most powerful AI training system in the world when it came online, and it has since grown into the anchor of a bigger fleet that now includes a second site, Colossus 2.

Why Colossus 1 Exists

xAI joined an AI race that was already crowded with well-funded rivals, and large language models mostly get better by training on more data, across more chips, for longer. To close that gap quickly, xAI needed a very large cluster, and it needed one fast. Building a data center from scratch can take years, but taking over a shuttered factory that already had heavy industrial power and a huge open floor let the team get running in 122 days. That is how Elon works: instead of accepting the pace everyone else assumes is normal, he attacks the physics and the logistics head-on and compresses the timeline.

How it Works

Colossus 1 works less like one machine and more like a tightly woven web of GPUs that all have to think as a single brain. Each server moves data at 3.6 terabits per second over NVIDIA Spectrum-X Ethernet, a standard networking approach xAI chose over more exotic options, and the full setup reaches a combined memory bandwidth of 194 petabytes per second, with storage past one exabyte, which is a billion gigabytes. Feeding all of that power is the hard part, because training loads swing hard as the chips fire and go quiet in unison. So xAI paired the site with Tesla Megapack batteries to smooth out the draw, which is a neat piece of Elon connecting his own companies, since the batteries come from Tesla.

The local grid had to grow to keep up. Power approved for the site from the Tennessee Valley Authority, delivered through Memphis Light, Gas and Water, started at 150 megawatts and doubled to 300 megawatts once a second approval cleared in early 2026.

The mix of chips changed as fast as the building filled up. When Elon laid out the jump to 200,000 GPUs, he was specific about it: the original H100s plus a batch of newer H200s, with later expansions adding GB200 Blackwell-class hardware on top.

The Economics of Colossus 1

Each new generation of chip costs more and does more work per dollar, and a cluster this size starts losing value the moment it turns on. So in May 2026 xAI made a smart move: it opened Colossus 1, by then more than 220,000 mixed H100, H200, and GB200 GPUs, to Anthropic, the maker of Claude, adding capacity for Claude Pro and Claude Max users. The older cluster became steady income while the newest training work moved to the Blackwell-based Colossus 2. The same chips that train Grok now also serve a direct competitor, which is one way to keep the machine busy and the margins high. The full build runs into tens of billions of dollars, including a roughly 20 billion dollar Mississippi expansion announced in January 2026.

Current Status of Colossus 1

The scale is what stands out most. Colossus 1 opened at 100,000 GPUs, roughly doubled within months, and by early 2026 the combined Colossus 1 and Colossus 2 fleet reached about 555,000 NVIDIA GPUs, with a stated plan to push toward one million. The clearest way to see that growth is in the power it uses. Colossus 1 alone now draws 300 megawatts of approved grid power, and the two clusters together pull roughly 2 gigawatts, so a single site now uses about as much electricity as a mid-sized city.

What Comes Next

Colossus 2 is xAI's claim to the first gigawatt-scale AI data center, and it is where Grok 5 is being trained on Blackwell GB200 and GB300 hardware. The January 2026 Mississippi expansion, codenamed MACROHARDrr, together with the two Colossus sites, is expected to form a roughly 2-gigawatt system with more than one million GPUs. The Anthropic deal even mentioned interest in multi-gigawatt orbital compute, which means data centers in space, a reminder that for Elon the ground is rarely the limit.

The Bottom Line

Colossus 1 took an empty appliance plant and turned it into the beating heart of xAI in 122 days, then kept growing until it became infrastructure that other AI labs would pay to use. The name fits, because this is raw scale used as strategy, and the bet is simple: whoever builds the biggest, fastest training engine gets to shape the models everyone else relies on.

Related

Keep reading: Colossus 2, Grok. Zoom out to the SpaceXAI overview, or open the Quote Library.