xAI plans to bring ~420,000 Nvidia GPUs online at its Colossus facility in November 2026, marking one of the largest single-site GPU deployments globally.
SpaceXAI plans to activate approximately 420,000 Nvidia GPUs in November to meet fresh compute commitments. The ramp is part of a broader buildout of Colossus, the supercomputer cluster in Memphis, Tennessee, that powers Elon Musk's AI ambitions. The timing is strategic: the company has begun leasing significant capacity to Anthropic, a direct rival in the model race.
**What is inside Colossus**
On September 25, 2026, Musk detailed the current hardware across the two Colossus sites.
Colossus 1 houses three Nvidia generations: 150,000 H100 chips, 50,000 H200 chips, and 30,000 GB200 chips, totaling 230,000 GPUs.
Colossus 2 is newer and significantly larger, containing 110,000 GB200 chips and 440,000 GB300 chips—550,000 in total.
Combined, the two facilities house approximately 780,000 Nvidia GPUs. The Colossus 2 figures reportedly reflect capacity limits tied to networking constraints.
**The path to 1.44 million**
Musk announced plans to bring 660,000 additional GPUs online by year-end. If achieved, the total fleet would reach approximately 1.44 million. The roughly 420,000 GPUs slated for November fit within that timeline, nearly doubling the cluster from 780,000 to 1.44 million in a single quarter.
The hardware will support training and inference for Grok, the company's family of AI models, including Grok 4.7. SpaceXAI is building a 1.2-gigawatt power plant to supply the expanded cluster.
**Leasing to a rival**
SpaceXAI has leased over 220,000 GPUs from Colossus 1 to Anthropic. The arrangement could be worth up to $84.5 billion by 2029—a potential ceiling tied to future commitments, not currently collected cash. Since Colossus 1 holds 230,000 GPUs in total, leasing over 220,000 means most of that site's capacity now trains someone else's models, explaining the urgency behind the November activation: new commitments require new hardware.
**The 2030 vision**
Musk's longer-term target extends far beyond Memphis. The goal for 2030 is infrastructure equivalent to 50 million H100-class GPUs. Musk has consistently emphasized close partnerships with Nvidia.
**Implications**
A potential $84.5 billion leasing relationship with Anthropic signals that frontier labs are willing to sign enormous long-dated contracts to secure capacity. The 1.2-gigawatt plant underscores that power has become as strategic as silicon. Key metrics to track: whether the November activation hits schedule, whether year-end capacity approaches 1.44 million, whether the power plant comes online in time, and whether the Anthropic arrangement grows toward its potential value or plateaus.