SpaceX AI plans Texas data center as CapEx surges to $7.7 billion

SpaceX AI is pushing deeper into the AI infrastructure arms race, scouting sites in Texas for at least one gigawatt-scale data center that could rival or exceed its Memphis facility in size.

SpaceX AI has evaluated multiple locations across Texas and laid groundwork for at least one large data center, according to The Information. The company has relocated existing data center employees to the region and posted several infrastructure-related job openings in Austin, Bastrop and the Corsicana area in recent weeks.

The expansion comes as SpaceX AI's capital spending accelerates sharply. The company spent $7.7 billion on AI-related capital expenditures in the first quarter of 2026, roughly three times the amount from a year earlier. Morgan Stanley analysts have projected SpaceX AI will burn through $120 billion in cash combined across 2026 and 2027, driven primarily by AI infrastructure investment.

SpaceX AI currently operates about 1 gigawatt of computing capacity across two active data centers, housing hundreds of thousands of Nvidia graphics processing units. Supermicro Chief Executive Charles Liang said in June on X that his company is helping SpaceX AI build "another brand new gigawatt-scale AI data center" within one year, though he did not disclose the location.

The Texas push follows the playbook SpaceX AI used in Memphis, where it converted an abandoned manufacturing facility into its first data center. The company is weighing similar conversions of existing warehouses in Texas alongside ground-up construction of multiple facilities, two people familiar with the plans told The Information. A timeline for the expansion has not been set.

Energy strategy draws scrutiny

SpaceX AI's reliance on natural gas turbines to supplement grid power for its data centers has become a flashpoint, particularly in Memphis. The company has installed dozens of mobile turbines across Tennessee and Mississippi and is developing a permanent gas-fired power plant in Mississippi through a joint venture with Houston-based Solaris Energy Infrastructure.

The approach has drawn opposition from local residents, politicians and environmental groups in Memphis over turbine use and water consumption. Tensions escalated earlier this year when SpaceX AI delayed construction of an $80 million wastewater recycling facility as part of a cost-cutting push.

In the Memphis region, SpaceX AI continues work on a third data center. It also purchased about 75,000 square meters of warehouse space in Mississippi late last year and has stationed employees in Jacksonville, Florida, where it posted data center job openings. The company's May acquisition of APR Energy, a power equipment supplier based in Jacksonville with operations across Tennessee, Mississippi and Texas, further signals its push to control its energy supply chain.

Texas as a strategic hub

Texas holds particular strategic importance for Elon Musk's business empire. SpaceX and Tesla are jointly building a semiconductor manufacturing facility in Grimes County near Corsicana, and SpaceX AI's data center expansion aligns with the broader geographic clustering of Musk's ventures in the state.

The expansion adds to SpaceX AI's already heavy capital burden. To offset spending, the company in June began leasing substantial computing capacity to Anthropic and Google, with Musk saying SpaceX AI can reclaim the capacity when needed. SpaceX AI also reached a $60 billion agreement to acquire AI startup Cursor, which is working to expand beyond coding tools and develop a leading general-purpose model by year-end.

The data center land grab is playing out across the industry. A consortium including BlackRock's Global Infrastructure Partners, MGX and the AI Infrastructure Partnership this week closed the largest data center acquisition on record, buying Aligned Data Centers for about $40 billion, then committing an additional $5 billion in growth capital. The deal underscores the premium investors are placing on AI-ready capacity with secured power and cooling infrastructure.

For investors, the question is whether the spending will pay off. SpaceX AI's projected $120 billion cash burn over two years dwarfs the combined capital budgets of most hyperscalers. Nvidia, whose GPUs power the bulk of SpaceX AI's infrastructure, stands to benefit from every new facility. Supermicro, which supplies servers and cooling equipment, has tied its growth trajectory directly to SpaceX AI's build-out. But if AI demand growth stalls, the billions committed to concrete, turbines and silicon could become one of the most expensive overhangs in technology history.

This article is for informational purposes only and does not constitute investment advice.