SpaceX targets 10GW AI computing expansion amid electricity and space satellite developments

SpaceX Pursues Aggressive AI Computing Infrastructure Expansion

SpaceX claims it could add 10 gigawatts of additional data center capacity during 2027, an ambition that would match what Google accumulated throughout its entire existence1

1

. Elon Musk's company plans to increase its AI computing infrastructure from approximately 3GW to more than 10GW within 18 months through developments planned across several locations. According to SemiAnalysis, this expansion schedule remains achievable, citing construction progress, internal manufacturing capabilities, and access to dedicated power generation equipment1

1

.

Colossus 2 Demonstrates SpaceX's Construction Capabilities

The rapid development of Colossus 1 and Colossus 2 provides evidence that Musk's ambitious construction schedule could become reality. Colossus 2 represents a particularly important example, with its electricity generation capacity reportedly increasing from 0.4GW in January toward 2.4GW by December1

1

. The facility includes 89 gas turbines capable of generating approximately 2.5GW, alongside 1,377 Tesla Megapack battery systems installed across the property. Musk previously stated that Colossus 2 already operated 550,000 GPUs as of September 251

1

.

Electricity Challenge Threatens AI Infrastructure Deployment

Environmental restrictions will eventually limit the permitted output from Colossus 2's turbines to approximately 1.24GW, creating additional requirements for electricity supply and presenting a significant electricity challenge1

1

. The facility is expected to receive approximately 1.2GW from the electricity grid, although analysts estimate another 300MW remains necessary. Battery storage could help address that difference by supplying additional electricity when demand fluctuates. SpaceX reportedly has temporary permission from Mississippi authorities to operate additional turbines, provided the company phases them out by July 20271

1

.

Tesla Manufacturing Capabilities Could Accelerate Construction

One factor potentially accelerating construction is SpaceX's relationship with Tesla, which could provide access to manufacturing capabilities and electrical infrastructure components. SemiAnalysis suggests Tesla may be producing modular mechanical and electrical equipment, allowing installation teams to deploy completed systems without waiting for conventional suppliers1

1

. Such an arrangement could reduce procurement delays affecting competing developers, particularly when specialized power generation equipment and other essential infrastructure remain difficult to obtain. The analysts identified approximately 9GW of potential compute capacity across the facilities included in their assessment of SpaceX's development pipeline1

1

.

Next-Generation AI Satellites Target Space Computing Market

SpaceX simultaneously pursues space AI computing through its Starmind plan, aiming to put data center satellites in orbit. Analyst Jeff Pu at GF Securities believes SpaceX should have 5 gigawatts of space AI computing capacity by 2028 courtesy of 10,000 NVIDIA GPU racks2

2

. The AI1 satellites are designed to carry compute payloads, which include NVIDIA's Vera Rubin platform. According to SpaceX, the AI1 model is capable of delivering a 250-kilowatt compute payload per satellite2

2

.

Advanced GPU Specifications for Space-Based Computing

Pu's estimate relies on what he dubs Kessel racks that use NVIDIA's Vera Rubin AI systems. The analyst believes that the thermal design point of each GPU present in the NVL72 platform for the Kessel racks will be 4,000 watts, a significant bump over what rumors have pointed towards as the power dissipation of standard terrestrial NVIDIA GPUs2

2

. The Kessel is a compact NVL72, with SpaceX and NVIDIA having spread the racks horizontally for a 9U configuration compared to a 48U configuration. A 9U rack is roughly 16 inches in height, considerably lower than 84 inches for a 48U configuration and necessary for payload stability during a rocket launch2

2

.

Regulatory Hurdles and Supply Chain Constraints Remain

Completed buildings do not automatically translate into operational AI computing infrastructure, particularly when electrical connections, cooling systems and server installations remain outstanding. Microsoft CEO Satya Nadella previously explained this problem, stating that chips sitting in inventory cannot be deployed without proper infrastructure1

1

. SpaceX must secure sufficient GPUs, memory components and power generation equipment before its ambitious deployment schedule can become reality. Even with substantial manufacturing resources, regulatory hurdles, construction materials and local opposition could complicate expansion across multiple locations1

1

. Grid limitations and environmental restrictions add further complexity to the timeline, making the 2027 target ambitious despite SpaceX's demonstrated construction capabilities.