Elon Musk Plans to Double Nvidia (NVDA) Chip Count at Colossus 2 Data Center by Year-End

Key Points

Table of Contents

xAI’s Colossus 2 data center in Memphis operates 110,000 Nvidia GB200 chips alongside 440,000 GB300 chips currently.
An additional 220,000 GB300 chips are scheduled to launch next week, according to Musk’s post on X.
The facility expects to receive another 220,000 GB300 chips in November.
A final batch of 220,000 chips may arrive by late December, pending schedule alignment, Musk indicated.
If all deployments proceed as planned, the facility’s chip inventory will approximately double before 2027.

Elon Musk provided fresh insights this week regarding the rapid expansion plans for his xAI business plans and its Colossus 2 computing facility. The data center operates in the Memphis, Tennessee region, positioned close to Southaven, Mississippi.

Elon Musk says $SPCX will have ~1 million GPUs online next week as another 220K GB300s come online at Colossus 2.

From there another 220K are planned for November and potentially 220K more by late December taking xAI to ~1.44 million GPUs by year end. pic.twitter.com/J70u0j8Z5r

— Shay Boloor (@StockSavvyShay) September 25, 2026

The Tesla CEO disclosed the information through a post on X platform in the early hours of Friday. His announcement indicated the facility could potentially increase its Nvidia chip deployment by more than double before the conclusion of 2026.

Based on Musk’s statement, the Colossus 2 facility presently operates with 110,000 Nvidia GB200 chips. Additionally, the center deploys 440,000 GB300 chips, representing Nvidia’s more recent hardware generation.

Phased Chip Deployment Through Year-End

The xAI chief executive announced that an initial wave of 220,000 GB300 chips will become operational within the next week. This deployment represents the opening phase of the broader expansion strategy.

Following the initial rollout, an additional 220,000 GB300 chips are slated for integration in November. Musk noted that a third installment of 220,000 chips might arrive before the year concludes, specifically in late December, although he cautioned this final batch remains contingent on scheduling.

Should the complete deployment schedule materialize as outlined, the Colossus 2 facility would conclude the year operating significantly more chips than its present configuration. Musk has not yet provided a consolidated total for the projected year-end chip count.

The SpaceX founder’s artificial intelligence venture previously announced its objective to outfit the Memphis location with 1 million graphics processing units. This recent announcement represents the most comprehensive roadmap Musk has shared toward achieving that ambitious target.

The announcement did not address whether the facility possesses adequate power infrastructure and available capacity to accommodate the incoming chips immediately. Installations of this magnitude demand substantial electrical supply and sophisticated cooling systems.

Colossus Facility Operations and Applications

Musk has characterized the Colossus infrastructure as the planet’s most powerful AI supercomputer. The inaugural installation, Colossus 1, launched in 2024 with the primary mission of training xAI’s Grok chatbot technology.

The facility’s purpose has expanded considerably beyond its original Grok training mandate. xAI has begun offering computing resources from the Memphis location to external organizations on a rental basis.

In recent months, xAI initiated rental agreements with Anthropic and Google, owned by Alphabet. This development means the Memphis supercomputer now powers artificial intelligence initiatives beyond xAI’s internal development efforts.

The aggressive expansion reflects ongoing strong demand for Nvidia’s AI-focused chips throughout the technology sector. Organizations developing large language models require extraordinary computing capabilities for training and operational deployment.

Both Nvidia’s GB200 and GB300 chips belong to the company’s Blackwell chip family. These processors are engineered specifically for demanding AI training workloads and advanced processing operations.

The velocity of Musk’s expansion strategy demonstrates how rapidly AI infrastructure initiatives are scaling throughout this year. Colossus 2 has evolved from its initial deployment to housing hundreds of thousands of processing units within a compressed timeframe.

Musk’s social media announcement did not detail the financial investment associated with acquiring and deploying the additional chips. The post also omitted mention of any new energy supply contracts supporting the expansion.

As of Friday’s update, the latest information from Musk confirms the initial deployment of 220,000 GB300 chips will activate next week, followed by the November installation and the potential December addition thereafter.

✨ Limited Time Offer