Elon Musk is preparing a major expansion of xAI's computing capacity, with plans for an AI supercomputer built around Nvidia graphics hardware. According to a report from The Information, which cites a presentation Musk gave to investors, the project is meant to give xAI far more compute for training and running advanced AI systems.
A new scale target for xAI
The planned system has a name that signals the ambition: the "Gigafactory of Compute." It is scheduled to be completed by the fall of 2025 and is projected to be at least four times larger than the most powerful clusters currently used by competitors such as Meta.
The machine is expected to use Nvidia H100 GPUs. Musk personally guarantees the timely completion of the supercomputer, according to the source article. One important detail remains unresolved: it is not yet clear where the cluster will be built.
For xAI, the project would move compute from a support function to a central strategic asset. Large AI models depend heavily on access to graphics processors, and the source frames the new system as an attempt to outbuild rival infrastructure rather than simply rent or assemble smaller capacity over time.
Why the comparison with Meta matters
Meta is the benchmark named in the report because of the size of its own GPU plans. Meta CEO Mark Zuckerberg said in January 2024 that Meta aims to have 340,000 Nvidia H100 GPUs and about 600,000 graphics cards in use by the end of the year.
Against that backdrop, xAI's proposed cluster is not just another data center project. The target described in the report would place the "Gigafactory of Compute" above the most powerful clusters currently used by competitors such as Meta, if it is completed as planned.
The comparison also shows how AI companies are now competing on infrastructure as much as on model design. The source does not describe the full technical layout of the supercomputer, but the emphasis on Nvidia H100 GPUs makes clear that access to advanced chips is central to the plan.
Oracle, Grok, and the current compute base
xAI may be partnering with Oracle on the new effort. Oracle already supports the training and inference of Grok, xAI's current AI chatbot on Musk's "X" platform.
According to The Information, xAI has an Oracle cluster of 16,000 H100 chips for Grok, and Grok is trained on a total of 20,000 GPUs. Nvidia has also reportedly given xAI priority status for shipping its new "Blackwell" AI GPU.
That current footprint gives useful context for the larger plan. xAI already has access to a significant GPU base for Grok, but the planned "Gigafactory of Compute" would represent a much larger push toward dedicated AI infrastructure.
Funding and model ambitions
The compute plan arrives as xAI is reportedly raising major capital. The company reportedly raised $6 billion in funding at the end of April at a pre-money valuation of $18 billion. The funding round is expected to close in the coming weeks.
The source also connects the infrastructure push to Grok's model roadmap. In early April, Musk announced a new Grok model coming in May that he claims is better than the GPT-4.
Musk also believes that AI capabilities will surpass human intelligence by the end of next year, given sufficient computing power and energy. That condition matters: in the account provided, compute and energy are not side issues, but requirements for the level of AI progress Musk is describing.
The wider supercomputer race
xAI is not the only company tied to very large AI infrastructure plans. Microsoft and OpenAI are also reportedly planning to build a new supercomputer called "Stargate," which could cost up to $100 billion and be fully developed by 2030.
The same report says a smaller supercomputer, known as "Phase 4," could be operational as early as 2026. The full expansion of that project depends on whether OpenAI can make significant progress in its AI research toward superintelligence.
Taken together, these plans point to a clear industry direction from the facts in the source: leading AI companies are trying to secure more chips, larger clusters, and longer-term infrastructure control. For xAI, the "Gigafactory of Compute" is the proposed leap from today's Grok compute base to a much larger competitive platform.