Elon Musk has announced plans to expand xAI's computing infrastructure sevenfold by the end of 2027, potentially creating one of the world's most powerful AI clusters capable of generating $300 to $500 billion in annual revenue. The expansion comes as the company faces legal headwinds, with a federal judge blocking xAI's attempt to relocate a privacy lawsuit to Texas.

What Is xAI Planning to Build by 2027?

Musk told SpaceX employees that xAI's data centers will increase power capacity from the current 1.4 gigawatts (GW) to 10 GW by the end of 2027. To put that in perspective, this would represent roughly 2.1 to 2.4 million specialized AI processors working in concert, far exceeding the combined computing power of all supercomputers currently ranked in the world's Top 500 list.

The revenue projections are substantial. Musk estimated that if xAI achieves a value of $30 to $50 per watt of compute capacity, the 10 GW cluster could generate between $300 billion and $500 billion in annual revenue by the end of 2027. This assumes xAI can sell access to its computing infrastructure to other AI companies and enterprises at competitive rates.

How Does This Computing Power Compare to Existing Systems?

The scale of xAI's planned expansion is difficult to overstate. Current projections suggest the cluster will deliver between 105 and 120 exaFLOPS of inference performance, a measure of how many floating-point calculations the system can perform per second. For training AI models, the cluster is expected to achieve 73 to 84 exaFLOPS. These numbers represent a dramatic leap beyond anything currently operating publicly.

To contextualize: the combined FP64 (64-bit floating-point) performance of all systems on the Top 500 supercomputer list totals just 18.73 exaFLOPS. xAI's planned cluster would deliver 3.7 to 4.3 times more raw computing power than that entire global ranking combined. This capability could enable training of far more sophisticated AI models than currently possible.

Steps to Understanding xAI's Infrastructure Roadmap

• Current Baseline: xAI's Memphis and Southaven data centers currently draw 1.4 GW of power, with approximately 0.8 GW to 0.95 GW of that power available to AI accelerators across a mix of Hopper and Blackwell GPU architectures.
• Expansion Target: The company plans to add approximately 8.6 GW of additional nameplate power capacity by late 2027, requiring massive investments in electrical infrastructure, cooling systems, and real estate to house the equipment.
• Hardware Shift: xAI is betting primarily on Nvidia's Vera Rubin GPU systems going forward, moving away from the current mix of older architectures to standardize on next-generation accelerators for better efficiency and performance.
• Revenue Model: If successful, the expanded cluster could generate $300 to $500 billion annually by selling compute access to other AI companies, though actual revenue will depend on market demand and competition from other providers.

What Legal Challenges Could Slow xAI's Growth?

While Musk's expansion plans are aggressive, xAI faces a privacy lawsuit that could distract management and create regulatory uncertainty. A federal judge ruled that xAI cannot move a user privacy suit to Texas, keeping the case in its original jurisdiction. The ruling suggests the company will need to defend itself in court rather than consolidate the case elsewhere.

Legal battles can be costly and time-consuming, particularly for a company in rapid growth mode. However, the judge's decision to block the transfer does not necessarily indicate the lawsuit has merit; it simply means the case will proceed where it was filed. The outcome could affect how xAI handles user data and privacy practices going forward, potentially requiring changes to how Grok, xAI's conversational AI assistant, collects and processes user information.

Why Should the AI Industry Pay Attention to This Expansion?

xAI's planned compute capacity would make it a major player in the AI infrastructure market, competing directly with established providers like Amazon Web Services, Microsoft Azure, and Google Cloud. The ability to offer 10 GW of cutting-edge compute capacity could attract customers developing large language models, training AI agents, and running inference workloads at scale.

However, Musk's revenue projections assume sustained demand and pricing power. The AI compute market is becoming increasingly competitive, with multiple companies racing to build massive data centers. Whether xAI can actually achieve $300 to $500 billion in annual revenue depends on factors beyond its control, including market saturation, customer acquisition costs, and the emergence of more efficient AI architectures that require less computing power.

The expansion also highlights the enormous capital requirements of modern AI infrastructure. Building and powering a 10 GW data center cluster requires not just billions in hardware investment, but also securing reliable electricity supplies and managing the thermal challenges of running millions of processors simultaneously. SpaceX's market capitalization has fluctuated significantly in recent months, dropping nearly $570 billion in less than two months before rebounding, which could affect the parent company's ability to fund xAI's ambitions.