Table of Contents
Elon Musk is planning a massive expansion of xAI’s data center infrastructure, with the company expected to increase its power capacity several times over by the end of 2027. Musk discussed the plan during a meeting with SpaceX employees, outlining an ambitious target for the company’s AI infrastructure.
According to Musk, xAI’s data centers could reach 10GW of power capacity by the end of 2027. That would be around seven times the company’s current capacity, which stands at roughly 1.4GW across its data centers in Memphis and Southaven.
10 gigawatts of AI compute will bring in billions of revenue, says Musk #
Musk also believes the expansion could translate into huge revenue. He estimated that 10GW of AI compute could potentially generate between $300 billion and $500 billion in annual revenue, based on his estimate of $30 to $50 in value for every watt of AI compute.
Latest PC & Tech deals
Acer Predator Helios Neo 16 - was $2,199 now $1,749
ASUS SFF-Ready RTX 5080 - was $1,499 now $1,354
Samsung Odyssey G7- was $999 now $649
HP OMEN 45L gaming PC - was $2,499 now $2,309
BIWIN Black Opal NV7400 4TB SSD - was $599 now $459
Prices correct as of July 10th, 2026.
The expansion would also significantly increase the amount of computing power available to xAI. The company already operates some of the world’s largest AI training clusters, and Musk expects its infrastructure at the end of next year to be around 10 times larger than what it has built so far.
However, power capacity should not be confused directly with computing performance. A data center’s total power draw also includes cooling systems, networking equipment, storage, CPUs, power delivery, and other infrastructure. Only part of the available power can actually be used by AI accelerators.
xAI is also expected to rely heavily on Nvidia’s next-generation hardware for the expansion. The company currently uses a combination of Nvidia Hopper and Blackwell accelerators, but Musk has indicated that xAI plans to use Nvidia’s upcoming Vera Rubin systems.
If the entire additional capacity were built around Nvidia’s expected NVL72 VR200 systems, the resulting infrastructure could potentially contain more than two million Rubin GPUs, based on estimates of power consumption and data center efficiency. That would make xAI’s planned infrastructure far larger than its current AI clusters. Still, reaching 10GW by the end of 2027 will require a huge amount of power, hardware, cooling infrastructure, and construction. It remains to be seen how quickly xAI can build out the capacity needed to meet Musk’s target.